{"license":"CC-BY-SA-4.0","licenseUrl":"https://creativecommons.org/licenses/by-sa/4.0/","attribution":"LiveBench 团队，livebench.ai，2026-06-25 题集，2026-09-28 网站材料。本站进行型号映射、类别均分、两位小数舍入与范围内排序；改编成绩数据沿用 CC BY-SA 4.0。未传播题目、回答或上游代码，不代表作者背书。","boards":[{"id":"livebench-overall","name":"LiveBench · 全科","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:78797b5914d0063bd3a43a98ce30b736dc3ffa4a6f11d850ce49caeec0055bef:1021e4ef211ccba609d6da2c22032f46d695c1546508cf144a58a2f3d6960f3d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-06-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 54 个型号 / 官方材料 66 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":74.64,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":73.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":66.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":69.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":77.97,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":80.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":62.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"grok-4.5","modelId":"model:xai/grok-4.5","label":"Grok 4.5","configuration":"Grok 4.5 · grok-4.5","value":75.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.5"]},{"id":"grok-build-0.1","modelId":"model:xai/grok-build-0.1","label":"Grok Build 0.1","configuration":"Grok Build 0.1 · grok-build-0.1","value":67.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-build-0.1"]},{"id":"kimi-k2.7-code","modelId":"model:moonshotai/kimi-k2.7-code","label":"Kimi K2.7 Code","configuration":"Kimi K2.7 Code · kimi-k2.7-code","value":68.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://moonshot.ai"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":64.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":68.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":73.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"kimi-k3","modelId":"model:moonshotai/kimi-k3","label":"Kimi K3","configuration":"Kimi K3 · kimi-k3","value":79.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k3"]},{"id":"deepseek-v4-flash-0731","modelId":"model:deepseek/deepseek-v4-flash-0731","label":"DeepSeek V4 Flash 0731","configuration":"DeepSeek V4 Flash 0731 · deepseek-v4-flash-0731","value":74.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/quick_start/pricing/"]},{"id":"qwen3.8-max","modelId":"model:alibaba/qwen3.8-max","label":"Qwen 3.8 Max","configuration":"Qwen 3.8 Max · qwen3.8-max","value":78.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"deepseek-v4-pro-0813","modelId":"model:deepseek/deepseek-v4-pro-0813","label":"DeepSeek V4 Pro 0813","configuration":"DeepSeek V4 Pro 0813 · deepseek-v4-pro-0813","value":77.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/news/news260424"]},{"id":"grok-4.6","modelId":"model:xai/grok-4.6","label":"Grok 4.6 xHigh","configuration":"Grok 4.6 xHigh · grok-4.6","value":78.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://x.ai/news/grok-4-6"]},{"id":"qwen3.8-27b","modelId":"model:alibaba/qwen3.8-27b","label":"Qwen3.8 27B","configuration":"Qwen3.8 27B · qwen3.8-27b","value":75.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.qwencloud.com/models/qwen3.8-27b"]},{"id":"deepseek-v4-flash-vision-exp","modelId":"model:deepseek/deepseek-v4-flash-vision-exp","label":"DeepSeek V4 Flash Vision Exp","configuration":"DeepSeek V4 Flash Vision Exp · deepseek-v4-flash-vision-exp","value":76.76,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"qwen3.8-flash-next","modelId":"model:alibaba/qwen3.8-flash-next","label":"Qwen 3.8 Flash Next","configuration":"Qwen 3.8 Flash Next · qwen3.8-flash-next","value":76.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":67.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":72.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":74.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":76.53,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":72.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-sonnet-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5","label":"Claude Sonnet 5 xHigh Effort","configuration":"Claude Sonnet 5 xHigh Effort · claude-sonnet-5-xhigh-effort","value":76.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-5"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":76.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"glm-5.2","modelId":"model:zhipuai/glm-5.2","label":"GLM-5.2","configuration":"GLM-5.2 · glm-5.2","value":73.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":70.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"gpt-5.6-sol-max","modelId":"model:openai/gpt-5.6-sol","label":"GPT-5.6 Sol Max Effort","configuration":"GPT-5.6 Sol Max Effort · gpt-5.6-sol-max","value":81.05,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-terra-max","modelId":"model:openai/gpt-5.6-terra","label":"GPT-5.6 Terra Max Effort","configuration":"GPT-5.6 Terra Max Effort · gpt-5.6-terra-max","value":77.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-luna-max","modelId":"model:openai/gpt-5.6-luna","label":"GPT-5.6 Luna Max Effort","configuration":"GPT-5.6 Luna Max Effort · gpt-5.6-luna-max","value":73.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"claude-fable-5-max-effort","modelId":"model:anthropic/claude-fable-5","label":"Claude Fable 5 Max Effort","configuration":"Claude Fable 5 Max Effort · claude-fable-5-max-effort","value":82.97,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5"]},{"id":"muse-spark-1.1-xhigh","modelId":"model:meta/muse-spark-1.1","label":"Muse Spark 1.1 xHigh Effort","configuration":"Muse Spark 1.1 xHigh Effort · muse-spark-1.1-xhigh","value":75.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"inkling-xhigh","modelId":"model:thinkingmachines/inkling","label":"Inkling xHigh Effort","configuration":"Inkling xHigh Effort · inkling-xhigh","value":71.92,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://thinkingmachines.ai/"]},{"id":"gemini-3.5-flash-lite-high","modelId":"model:google/gemini-3.5-flash-lite","label":"Gemini 3.5 Flash-Lite High","configuration":"Gemini 3.5 Flash-Lite High · gemini-3.5-flash-lite-high","value":63.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite"]},{"id":"gemini-3.6-flash-high","modelId":"model:google/gemini-3.6-flash","label":"Gemini 3.6 Flash High","configuration":"Gemini 3.6 Flash High · gemini-3.6-flash-high","value":73.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"]},{"id":"claude-opus-5-max-effort","modelId":"model:anthropic/claude-opus-5","label":"Claude 5 Opus Thinking Max Effort","configuration":"Claude 5 Opus Thinking Max Effort · claude-opus-5-max-effort","value":80.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-5"]},{"id":"claude-opus-4-8-max-effort","modelId":"model:anthropic/claude-opus-4-8","label":"Claude 4.8 Opus Thinking Max Effort","configuration":"Claude 4.8 Opus Thinking Max Effort · claude-opus-4-8-max-effort","value":76.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-8"]},{"id":"muse-spark-1.2-xhigh","modelId":"model:meta/muse-spark-1.2","label":"Muse Spark 1.2 xHigh Effort","configuration":"Muse Spark 1.2 xHigh Effort · muse-spark-1.2-xhigh","value":77.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"gemini-3.7-flash-high","modelId":"model:google/gemini-3.7-flash","label":"Gemini 3.7 Flash High","configuration":"Gemini 3.7 Flash High · gemini-3.7-flash-high","value":78.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash"]},{"id":"glm-5.3","modelId":"model:zhipuai/glm-5.3","label":"GLM-5.3","configuration":"GLM-5.3 · glm-5.3","value":76.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"glm-5.3-flash","modelId":"model:zhipuai/glm-5.3-flash","label":"GLM-5.3 Flash","configuration":"GLM-5.3 Flash · glm-5.3-flash","value":71.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"claude-fable-5-1-max-effort","modelId":"model:anthropic/claude-fable-5-1","label":"Claude Fable 5.1 Max Effort","configuration":"Claude Fable 5.1 Max Effort · claude-fable-5-1-max-effort","value":83.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5-1"]},{"id":"gemini-3.8-flash-high","modelId":"model:google/gemini-3.8-flash","label":"Gemini 3.8 Flash High","configuration":"Gemini 3.8 Flash High · gemini-3.8-flash-high","value":75.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash"]},{"id":"muse-spark-1.3-xhigh","modelId":"model:meta/muse-spark-1.3","label":"Muse Spark 1.3 xHigh Effort","configuration":"Muse Spark 1.3 xHigh Effort · muse-spark-1.3-xhigh","value":81.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://dev.meta.ai/docs/models"]},{"id":"gpt-6-astra-max","modelId":"model:openai/gpt-6-astra","label":"GPT-6 Astra Max Effort","configuration":"GPT-6 Astra Max Effort · gpt-6-astra-max","value":82.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-astra"]},{"id":"deepseek-v4.1-flash-max","modelId":"model:deepseek/deepseek-v4.1-flash","label":"DeepSeek V4.1 Flash Max Effort","configuration":"DeepSeek V4.1 Flash Max Effort · deepseek-v4.1-flash-max","value":81.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"grok-4.7-xhigh","modelId":"model:xai/grok-4.7","label":"Grok 4.7 xHigh","configuration":"Grok 4.7 xHigh · grok-4.7-xhigh","value":77.4,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.7"]},{"id":"claude-opus-5-5-max-effort","modelId":"model:anthropic/claude-opus-5-5","label":"Claude 5.5 Opus Thinking Max Effort","configuration":"Claude 5.5 Opus Thinking Max Effort · claude-opus-5-5-max-effort","value":83.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/opus-5-5/overview"]},{"id":"gpt-6-sol-max","modelId":"model:openai/gpt-6-sol","label":"GPT-6 Sol Max Effort","configuration":"GPT-6 Sol Max Effort · gpt-6-sol-max","value":79.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-sol"]},{"id":"gpt-6-luna-max","modelId":"model:openai/gpt-6-luna","label":"GPT-6 Luna Max Effort","configuration":"GPT-6 Luna Max Effort · gpt-6-luna-max","value":72.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-luna"]},{"id":"claude-sonnet-5-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5-5","label":"Claude Sonnet 5.5 xHigh Effort","configuration":"Claude Sonnet 5.5 xHigh Effort · claude-sonnet-5-5-xhigh-effort","value":77.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/sonnet-5-5/overview"]}]},{"id":"livebench-reasoning","name":"LiveBench · 推理","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:78797b5914d0063bd3a43a98ce30b736dc3ffa4a6f11d850ce49caeec0055bef:1021e4ef211ccba609d6da2c22032f46d695c1546508cf144a58a2f3d6960f3d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-06-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 54 个型号 / 官方材料 66 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":77.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":71.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":81.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":88.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":89.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":70.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"grok-4.5","modelId":"model:xai/grok-4.5","label":"Grok 4.5","configuration":"Grok 4.5 · grok-4.5","value":87.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.5"]},{"id":"grok-build-0.1","modelId":"model:xai/grok-build-0.1","label":"Grok Build 0.1","configuration":"Grok Build 0.1 · grok-build-0.1","value":76.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-build-0.1"]},{"id":"kimi-k2.7-code","modelId":"model:moonshotai/kimi-k2.7-code","label":"Kimi K2.7 Code","configuration":"Kimi K2.7 Code · kimi-k2.7-code","value":82.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://moonshot.ai"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":70.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":75.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":83.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"kimi-k3","modelId":"model:moonshotai/kimi-k3","label":"Kimi K3","configuration":"Kimi K3 · kimi-k3","value":90.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k3"]},{"id":"deepseek-v4-flash-0731","modelId":"model:deepseek/deepseek-v4-flash-0731","label":"DeepSeek V4 Flash 0731","configuration":"DeepSeek V4 Flash 0731 · deepseek-v4-flash-0731","value":86.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/quick_start/pricing/"]},{"id":"qwen3.8-max","modelId":"model:alibaba/qwen3.8-max","label":"Qwen 3.8 Max","configuration":"Qwen 3.8 Max · qwen3.8-max","value":88.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"deepseek-v4-pro-0813","modelId":"model:deepseek/deepseek-v4-pro-0813","label":"DeepSeek V4 Pro 0813","configuration":"DeepSeek V4 Pro 0813 · deepseek-v4-pro-0813","value":85.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/news/news260424"]},{"id":"grok-4.6","modelId":"model:xai/grok-4.6","label":"Grok 4.6 xHigh","configuration":"Grok 4.6 xHigh · grok-4.6","value":90.51,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://x.ai/news/grok-4-6"]},{"id":"qwen3.8-27b","modelId":"model:alibaba/qwen3.8-27b","label":"Qwen3.8 27B","configuration":"Qwen3.8 27B · qwen3.8-27b","value":80.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.qwencloud.com/models/qwen3.8-27b"]},{"id":"deepseek-v4-flash-vision-exp","modelId":"model:deepseek/deepseek-v4-flash-vision-exp","label":"DeepSeek V4 Flash Vision Exp","configuration":"DeepSeek V4 Flash Vision Exp · deepseek-v4-flash-vision-exp","value":85.4,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"qwen3.8-flash-next","modelId":"model:alibaba/qwen3.8-flash-next","label":"Qwen 3.8 Flash Next","configuration":"Qwen 3.8 Flash Next · qwen3.8-flash-next","value":87.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":74.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":80.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":88.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":87.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":84.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-sonnet-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5","label":"Claude Sonnet 5 xHigh Effort","configuration":"Claude Sonnet 5 xHigh Effort · claude-sonnet-5-xhigh-effort","value":88.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-5"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"glm-5.2","modelId":"model:zhipuai/glm-5.2","label":"GLM-5.2","configuration":"GLM-5.2 · glm-5.2","value":78.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":79.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"gpt-5.6-sol-max","modelId":"model:openai/gpt-5.6-sol","label":"GPT-5.6 Sol Max Effort","configuration":"GPT-5.6 Sol Max Effort · gpt-5.6-sol-max","value":91.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-terra-max","modelId":"model:openai/gpt-5.6-terra","label":"GPT-5.6 Terra Max Effort","configuration":"GPT-5.6 Terra Max Effort · gpt-5.6-terra-max","value":90.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-luna-max","modelId":"model:openai/gpt-5.6-luna","label":"GPT-5.6 Luna Max Effort","configuration":"GPT-5.6 Luna Max Effort · gpt-5.6-luna-max","value":85.64,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"claude-fable-5-max-effort","modelId":"model:anthropic/claude-fable-5","label":"Claude Fable 5 Max Effort","configuration":"Claude Fable 5 Max Effort · claude-fable-5-max-effort","value":89.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5"]},{"id":"muse-spark-1.1-xhigh","modelId":"model:meta/muse-spark-1.1","label":"Muse Spark 1.1 xHigh Effort","configuration":"Muse Spark 1.1 xHigh Effort · muse-spark-1.1-xhigh","value":87.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"inkling-xhigh","modelId":"model:thinkingmachines/inkling","label":"Inkling xHigh Effort","configuration":"Inkling xHigh Effort · inkling-xhigh","value":78.35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://thinkingmachines.ai/"]},{"id":"gemini-3.5-flash-lite-high","modelId":"model:google/gemini-3.5-flash-lite","label":"Gemini 3.5 Flash-Lite High","configuration":"Gemini 3.5 Flash-Lite High · gemini-3.5-flash-lite-high","value":60.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite"]},{"id":"gemini-3.6-flash-high","modelId":"model:google/gemini-3.6-flash","label":"Gemini 3.6 Flash High","configuration":"Gemini 3.6 Flash High · gemini-3.6-flash-high","value":85.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"]},{"id":"claude-opus-5-max-effort","modelId":"model:anthropic/claude-opus-5","label":"Claude 5 Opus Thinking Max Effort","configuration":"Claude 5 Opus Thinking Max Effort · claude-opus-5-max-effort","value":91.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-5"]},{"id":"claude-opus-4-8-max-effort","modelId":"model:anthropic/claude-opus-4-8","label":"Claude 4.8 Opus Thinking Max Effort","configuration":"Claude 4.8 Opus Thinking Max Effort · claude-opus-4-8-max-effort","value":89.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-8"]},{"id":"muse-spark-1.2-xhigh","modelId":"model:meta/muse-spark-1.2","label":"Muse Spark 1.2 xHigh Effort","configuration":"Muse Spark 1.2 xHigh Effort · muse-spark-1.2-xhigh","value":90,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"gemini-3.7-flash-high","modelId":"model:google/gemini-3.7-flash","label":"Gemini 3.7 Flash High","configuration":"Gemini 3.7 Flash High · gemini-3.7-flash-high","value":87.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash"]},{"id":"glm-5.3","modelId":"model:zhipuai/glm-5.3","label":"GLM-5.3","configuration":"GLM-5.3 · glm-5.3","value":85.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"glm-5.3-flash","modelId":"model:zhipuai/glm-5.3-flash","label":"GLM-5.3 Flash","configuration":"GLM-5.3 Flash · glm-5.3-flash","value":77.64,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"claude-fable-5-1-max-effort","modelId":"model:anthropic/claude-fable-5-1","label":"Claude Fable 5.1 Max Effort","configuration":"Claude Fable 5.1 Max Effort · claude-fable-5-1-max-effort","value":91.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5-1"]},{"id":"gemini-3.8-flash-high","modelId":"model:google/gemini-3.8-flash","label":"Gemini 3.8 Flash High","configuration":"Gemini 3.8 Flash High · gemini-3.8-flash-high","value":89.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash"]},{"id":"muse-spark-1.3-xhigh","modelId":"model:meta/muse-spark-1.3","label":"Muse Spark 1.3 xHigh Effort","configuration":"Muse Spark 1.3 xHigh Effort · muse-spark-1.3-xhigh","value":89.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://dev.meta.ai/docs/models"]},{"id":"gpt-6-astra-max","modelId":"model:openai/gpt-6-astra","label":"GPT-6 Astra Max Effort","configuration":"GPT-6 Astra Max Effort · gpt-6-astra-max","value":92.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-astra"]},{"id":"deepseek-v4.1-flash-max","modelId":"model:deepseek/deepseek-v4.1-flash","label":"DeepSeek V4.1 Flash Max Effort","configuration":"DeepSeek V4.1 Flash Max Effort · deepseek-v4.1-flash-max","value":86.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"grok-4.7-xhigh","modelId":"model:xai/grok-4.7","label":"Grok 4.7 xHigh","configuration":"Grok 4.7 xHigh · grok-4.7-xhigh","value":82.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.7"]},{"id":"claude-opus-5-5-max-effort","modelId":"model:anthropic/claude-opus-5-5","label":"Claude 5.5 Opus Thinking Max Effort","configuration":"Claude 5.5 Opus Thinking Max Effort · claude-opus-5-5-max-effort","value":92.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/opus-5-5/overview"]},{"id":"gpt-6-sol-max","modelId":"model:openai/gpt-6-sol","label":"GPT-6 Sol Max Effort","configuration":"GPT-6 Sol Max Effort · gpt-6-sol-max","value":88.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-sol"]},{"id":"gpt-6-luna-max","modelId":"model:openai/gpt-6-luna","label":"GPT-6 Luna Max Effort","configuration":"GPT-6 Luna Max Effort · gpt-6-luna-max","value":81.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-luna"]},{"id":"claude-sonnet-5-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5-5","label":"Claude Sonnet 5.5 xHigh Effort","configuration":"Claude Sonnet 5.5 xHigh Effort · claude-sonnet-5-5-xhigh-effort","value":86.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/sonnet-5-5/overview"]}]},{"id":"livebench-coding","name":"LiveBench · 编程答题","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:78797b5914d0063bd3a43a98ce30b736dc3ffa4a6f11d850ce49caeec0055bef:1021e4ef211ccba609d6da2c22032f46d695c1546508cf144a58a2f3d6960f3d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-06-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 54 个型号 / 官方材料 66 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":78.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":83.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":71.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":70.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":77.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":82.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":69.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"grok-4.5","modelId":"model:xai/grok-4.5","label":"Grok 4.5","configuration":"Grok 4.5 · grok-4.5","value":68.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.5"]},{"id":"grok-build-0.1","modelId":"model:xai/grok-build-0.1","label":"Grok Build 0.1","configuration":"Grok Build 0.1 · grok-build-0.1","value":65.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-build-0.1"]},{"id":"kimi-k2.7-code","modelId":"model:moonshotai/kimi-k2.7-code","label":"Kimi K2.7 Code","configuration":"Kimi K2.7 Code · kimi-k2.7-code","value":73.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://moonshot.ai"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":71.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":78.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":74.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"kimi-k3","modelId":"model:moonshotai/kimi-k3","label":"Kimi K3","configuration":"Kimi K3 · kimi-k3","value":81.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k3"]},{"id":"deepseek-v4-flash-0731","modelId":"model:deepseek/deepseek-v4-flash-0731","label":"DeepSeek V4 Flash 0731","configuration":"DeepSeek V4 Flash 0731 · deepseek-v4-flash-0731","value":74.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/quick_start/pricing/"]},{"id":"qwen3.8-max","modelId":"model:alibaba/qwen3.8-max","label":"Qwen 3.8 Max","configuration":"Qwen 3.8 Max · qwen3.8-max","value":72.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"deepseek-v4-pro-0813","modelId":"model:deepseek/deepseek-v4-pro-0813","label":"DeepSeek V4 Pro 0813","configuration":"DeepSeek V4 Pro 0813 · deepseek-v4-pro-0813","value":77.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/news/news260424"]},{"id":"grok-4.6","modelId":"model:xai/grok-4.6","label":"Grok 4.6 xHigh","configuration":"Grok 4.6 xHigh · grok-4.6","value":76.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://x.ai/news/grok-4-6"]},{"id":"qwen3.8-27b","modelId":"model:alibaba/qwen3.8-27b","label":"Qwen3.8 27B","configuration":"Qwen3.8 27B · qwen3.8-27b","value":75.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.qwencloud.com/models/qwen3.8-27b"]},{"id":"deepseek-v4-flash-vision-exp","modelId":"model:deepseek/deepseek-v4-flash-vision-exp","label":"DeepSeek V4 Flash Vision Exp","configuration":"DeepSeek V4 Flash Vision Exp · deepseek-v4-flash-vision-exp","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"qwen3.8-flash-next","modelId":"model:alibaba/qwen3.8-flash-next","label":"Qwen 3.8 Flash Next","configuration":"Qwen 3.8 Flash Next · qwen3.8-flash-next","value":72.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":70.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":79.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":78.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":82.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":79.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-sonnet-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5","label":"Claude Sonnet 5 xHigh Effort","configuration":"Claude Sonnet 5 xHigh Effort · claude-sonnet-5-xhigh-effort","value":80.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-5"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":76.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"glm-5.2","modelId":"model:zhipuai/glm-5.2","label":"GLM-5.2","configuration":"GLM-5.2 · glm-5.2","value":79.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":78.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"gpt-5.6-sol-max","modelId":"model:openai/gpt-5.6-sol","label":"GPT-5.6 Sol Max Effort","configuration":"GPT-5.6 Sol Max Effort · gpt-5.6-sol-max","value":83.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-terra-max","modelId":"model:openai/gpt-5.6-terra","label":"GPT-5.6 Terra Max Effort","configuration":"GPT-5.6 Terra Max Effort · gpt-5.6-terra-max","value":78.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-luna-max","modelId":"model:openai/gpt-5.6-luna","label":"GPT-5.6 Luna Max Effort","configuration":"GPT-5.6 Luna Max Effort · gpt-5.6-luna-max","value":82.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"claude-fable-5-max-effort","modelId":"model:anthropic/claude-fable-5","label":"Claude Fable 5 Max Effort","configuration":"Claude Fable 5 Max Effort · claude-fable-5-max-effort","value":85.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5"]},{"id":"muse-spark-1.1-xhigh","modelId":"model:meta/muse-spark-1.1","label":"Muse Spark 1.1 xHigh Effort","configuration":"Muse Spark 1.1 xHigh Effort · muse-spark-1.1-xhigh","value":77.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"inkling-xhigh","modelId":"model:thinkingmachines/inkling","label":"Inkling xHigh Effort","configuration":"Inkling xHigh Effort · inkling-xhigh","value":71.02,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://thinkingmachines.ai/"]},{"id":"gemini-3.5-flash-lite-high","modelId":"model:google/gemini-3.5-flash-lite","label":"Gemini 3.5 Flash-Lite High","configuration":"Gemini 3.5 Flash-Lite High · gemini-3.5-flash-lite-high","value":76.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite"]},{"id":"gemini-3.6-flash-high","modelId":"model:google/gemini-3.6-flash","label":"Gemini 3.6 Flash High","configuration":"Gemini 3.6 Flash High · gemini-3.6-flash-high","value":77.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"]},{"id":"claude-opus-5-max-effort","modelId":"model:anthropic/claude-opus-5","label":"Claude 5 Opus Thinking Max Effort","configuration":"Claude 5 Opus Thinking Max Effort · claude-opus-5-max-effort","value":81.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-5"]},{"id":"claude-opus-4-8-max-effort","modelId":"model:anthropic/claude-opus-4-8","label":"Claude 4.8 Opus Thinking Max Effort","configuration":"Claude 4.8 Opus Thinking Max Effort · claude-opus-4-8-max-effort","value":81.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-8"]},{"id":"muse-spark-1.2-xhigh","modelId":"model:meta/muse-spark-1.2","label":"Muse Spark 1.2 xHigh Effort","configuration":"Muse Spark 1.2 xHigh Effort · muse-spark-1.2-xhigh","value":77.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"gemini-3.7-flash-high","modelId":"model:google/gemini-3.7-flash","label":"Gemini 3.7 Flash High","configuration":"Gemini 3.7 Flash High · gemini-3.7-flash-high","value":78.89,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash"]},{"id":"glm-5.3","modelId":"model:zhipuai/glm-5.3","label":"GLM-5.3","configuration":"GLM-5.3 · glm-5.3","value":78.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"glm-5.3-flash","modelId":"model:zhipuai/glm-5.3-flash","label":"GLM-5.3 Flash","configuration":"GLM-5.3 Flash · glm-5.3-flash","value":78.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"claude-fable-5-1-max-effort","modelId":"model:anthropic/claude-fable-5-1","label":"Claude Fable 5.1 Max Effort","configuration":"Claude Fable 5.1 Max Effort · claude-fable-5-1-max-effort","value":86.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5-1"]},{"id":"gemini-3.8-flash-high","modelId":"model:google/gemini-3.8-flash","label":"Gemini 3.8 Flash High","configuration":"Gemini 3.8 Flash High · gemini-3.8-flash-high","value":72.49,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash"]},{"id":"muse-spark-1.3-xhigh","modelId":"model:meta/muse-spark-1.3","label":"Muse Spark 1.3 xHigh Effort","configuration":"Muse Spark 1.3 xHigh Effort · muse-spark-1.3-xhigh","value":81.06,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://dev.meta.ai/docs/models"]},{"id":"gpt-6-astra-max","modelId":"model:openai/gpt-6-astra","label":"GPT-6 Astra Max Effort","configuration":"GPT-6 Astra Max Effort · gpt-6-astra-max","value":80.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-astra"]},{"id":"deepseek-v4.1-flash-max","modelId":"model:deepseek/deepseek-v4.1-flash","label":"DeepSeek V4.1 Flash Max Effort","configuration":"DeepSeek V4.1 Flash Max Effort · deepseek-v4.1-flash-max","value":80.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"grok-4.7-xhigh","modelId":"model:xai/grok-4.7","label":"Grok 4.7 xHigh","configuration":"Grok 4.7 xHigh · grok-4.7-xhigh","value":77.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.7"]},{"id":"claude-opus-5-5-max-effort","modelId":"model:anthropic/claude-opus-5-5","label":"Claude 5.5 Opus Thinking Max Effort","configuration":"Claude 5.5 Opus Thinking Max Effort · claude-opus-5-5-max-effort","value":89.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/opus-5-5/overview"]},{"id":"gpt-6-sol-max","modelId":"model:openai/gpt-6-sol","label":"GPT-6 Sol Max Effort","configuration":"GPT-6 Sol Max Effort · gpt-6-sol-max","value":81.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-sol"]},{"id":"gpt-6-luna-max","modelId":"model:openai/gpt-6-luna","label":"GPT-6 Luna Max Effort","configuration":"GPT-6 Luna Max Effort · gpt-6-luna-max","value":78.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-luna"]},{"id":"claude-sonnet-5-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5-5","label":"Claude Sonnet 5.5 xHigh Effort","configuration":"Claude Sonnet 5.5 xHigh Effort · claude-sonnet-5-5-xhigh-effort","value":88.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/sonnet-5-5/overview"]}]},{"id":"livebench-agentic","name":"LiveBench · 代理编程","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:78797b5914d0063bd3a43a98ce30b736dc3ffa4a6f11d850ce49caeec0055bef:1021e4ef211ccba609d6da2c22032f46d695c1546508cf144a58a2f3d6960f3d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-06-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 54 个型号 / 官方材料 66 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":48.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":49.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":41.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":46.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":53.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":53.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":18.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"grok-4.5","modelId":"model:xai/grok-4.5","label":"Grok 4.5","configuration":"Grok 4.5 · grok-4.5","value":56.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.5"]},{"id":"grok-build-0.1","modelId":"model:xai/grok-build-0.1","label":"Grok Build 0.1","configuration":"Grok Build 0.1 · grok-build-0.1","value":45.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-build-0.1"]},{"id":"kimi-k2.7-code","modelId":"model:moonshotai/kimi-k2.7-code","label":"Kimi K2.7 Code","configuration":"Kimi K2.7 Code · kimi-k2.7-code","value":45.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://moonshot.ai"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":39.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":41.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":43.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"kimi-k3","modelId":"model:moonshotai/kimi-k3","label":"Kimi K3","configuration":"Kimi K3 · kimi-k3","value":62.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k3"]},{"id":"deepseek-v4-flash-0731","modelId":"model:deepseek/deepseek-v4-flash-0731","label":"DeepSeek V4 Flash 0731","configuration":"DeepSeek V4 Flash 0731 · deepseek-v4-flash-0731","value":46.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/quick_start/pricing/"]},{"id":"qwen3.8-max","modelId":"model:alibaba/qwen3.8-max","label":"Qwen 3.8 Max","configuration":"Qwen 3.8 Max · qwen3.8-max","value":64.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"deepseek-v4-pro-0813","modelId":"model:deepseek/deepseek-v4-pro-0813","label":"DeepSeek V4 Pro 0813","configuration":"DeepSeek V4 Pro 0813 · deepseek-v4-pro-0813","value":54.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/news/news260424"]},{"id":"grok-4.6","modelId":"model:xai/grok-4.6","label":"Grok 4.6 xHigh","configuration":"Grok 4.6 xHigh · grok-4.6","value":57.02,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://x.ai/news/grok-4-6"]},{"id":"qwen3.8-27b","modelId":"model:alibaba/qwen3.8-27b","label":"Qwen3.8 27B","configuration":"Qwen3.8 27B · qwen3.8-27b","value":61.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.qwencloud.com/models/qwen3.8-27b"]},{"id":"deepseek-v4-flash-vision-exp","modelId":"model:deepseek/deepseek-v4-flash-vision-exp","label":"DeepSeek V4 Flash Vision Exp","configuration":"DeepSeek V4 Flash Vision Exp · deepseek-v4-flash-vision-exp","value":65.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"qwen3.8-flash-next","modelId":"model:alibaba/qwen3.8-flash-next","label":"Qwen 3.8 Flash Next","configuration":"Qwen 3.8 Flash Next · qwen3.8-flash-next","value":61.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":38.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":39.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":48.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":50.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":42.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-sonnet-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5","label":"Claude Sonnet 5 xHigh Effort","configuration":"Claude Sonnet 5 xHigh Effort · claude-sonnet-5-xhigh-effort","value":59.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-5"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":44.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"glm-5.2","modelId":"model:zhipuai/glm-5.2","label":"GLM-5.2","configuration":"GLM-5.2 · glm-5.2","value":51.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":46.92,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"gpt-5.6-sol-max","modelId":"model:openai/gpt-5.6-sol","label":"GPT-5.6 Sol Max Effort","configuration":"GPT-5.6 Sol Max Effort · gpt-5.6-sol-max","value":56.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-terra-max","modelId":"model:openai/gpt-5.6-terra","label":"GPT-5.6 Terra Max Effort","configuration":"GPT-5.6 Terra Max Effort · gpt-5.6-terra-max","value":54.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-luna-max","modelId":"model:openai/gpt-5.6-luna","label":"GPT-5.6 Luna Max Effort","configuration":"GPT-5.6 Luna Max Effort · gpt-5.6-luna-max","value":48.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"claude-fable-5-max-effort","modelId":"model:anthropic/claude-fable-5","label":"Claude Fable 5 Max Effort","configuration":"Claude Fable 5 Max Effort · claude-fable-5-max-effort","value":62.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5"]},{"id":"muse-spark-1.1-xhigh","modelId":"model:meta/muse-spark-1.1","label":"Muse Spark 1.1 xHigh Effort","configuration":"Muse Spark 1.1 xHigh Effort · muse-spark-1.1-xhigh","value":58.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"inkling-xhigh","modelId":"model:thinkingmachines/inkling","label":"Inkling xHigh Effort","configuration":"Inkling xHigh Effort · inkling-xhigh","value":49.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://thinkingmachines.ai/"]},{"id":"gemini-3.5-flash-lite-high","modelId":"model:google/gemini-3.5-flash-lite","label":"Gemini 3.5 Flash-Lite High","configuration":"Gemini 3.5 Flash-Lite High · gemini-3.5-flash-lite-high","value":45.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite"]},{"id":"gemini-3.6-flash-high","modelId":"model:google/gemini-3.6-flash","label":"Gemini 3.6 Flash High","configuration":"Gemini 3.6 Flash High · gemini-3.6-flash-high","value":43.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"]},{"id":"claude-opus-5-max-effort","modelId":"model:anthropic/claude-opus-5","label":"Claude 5 Opus Thinking Max Effort","configuration":"Claude 5 Opus Thinking Max Effort · claude-opus-5-max-effort","value":65.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-5"]},{"id":"claude-opus-4-8-max-effort","modelId":"model:anthropic/claude-opus-4-8","label":"Claude 4.8 Opus Thinking Max Effort","configuration":"Claude 4.8 Opus Thinking Max Effort · claude-opus-4-8-max-effort","value":50.5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-8"]},{"id":"muse-spark-1.2-xhigh","modelId":"model:meta/muse-spark-1.2","label":"Muse Spark 1.2 xHigh Effort","configuration":"Muse Spark 1.2 xHigh Effort · muse-spark-1.2-xhigh","value":57.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"gemini-3.7-flash-high","modelId":"model:google/gemini-3.7-flash","label":"Gemini 3.7 Flash High","configuration":"Gemini 3.7 Flash High · gemini-3.7-flash-high","value":58.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash"]},{"id":"glm-5.3","modelId":"model:zhipuai/glm-5.3","label":"GLM-5.3","configuration":"GLM-5.3 · glm-5.3","value":60.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"glm-5.3-flash","modelId":"model:zhipuai/glm-5.3-flash","label":"GLM-5.3 Flash","configuration":"GLM-5.3 Flash · glm-5.3-flash","value":56.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"claude-fable-5-1-max-effort","modelId":"model:anthropic/claude-fable-5-1","label":"Claude Fable 5.1 Max Effort","configuration":"Claude Fable 5.1 Max Effort · claude-fable-5-1-max-effort","value":66.06,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5-1"]},{"id":"gemini-3.8-flash-high","modelId":"model:google/gemini-3.8-flash","label":"Gemini 3.8 Flash High","configuration":"Gemini 3.8 Flash High · gemini-3.8-flash-high","value":54.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash"]},{"id":"muse-spark-1.3-xhigh","modelId":"model:meta/muse-spark-1.3","label":"Muse Spark 1.3 xHigh Effort","configuration":"Muse Spark 1.3 xHigh Effort · muse-spark-1.3-xhigh","value":64.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://dev.meta.ai/docs/models"]},{"id":"gpt-6-astra-max","modelId":"model:openai/gpt-6-astra","label":"GPT-6 Astra Max Effort","configuration":"GPT-6 Astra Max Effort · gpt-6-astra-max","value":57.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-astra"]},{"id":"deepseek-v4.1-flash-max","modelId":"model:deepseek/deepseek-v4.1-flash","label":"DeepSeek V4.1 Flash Max Effort","configuration":"DeepSeek V4.1 Flash Max Effort · deepseek-v4.1-flash-max","value":77.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"grok-4.7-xhigh","modelId":"model:xai/grok-4.7","label":"Grok 4.7 xHigh","configuration":"Grok 4.7 xHigh · grok-4.7-xhigh","value":53.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.7"]},{"id":"claude-opus-5-5-max-effort","modelId":"model:anthropic/claude-opus-5-5","label":"Claude 5.5 Opus Thinking Max Effort","configuration":"Claude 5.5 Opus Thinking Max Effort · claude-opus-5-5-max-effort","value":71.72,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/opus-5-5/overview"]},{"id":"gpt-6-sol-max","modelId":"model:openai/gpt-6-sol","label":"GPT-6 Sol Max Effort","configuration":"GPT-6 Sol Max Effort · gpt-6-sol-max","value":52.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-sol"]},{"id":"gpt-6-luna-max","modelId":"model:openai/gpt-6-luna","label":"GPT-6 Luna Max Effort","configuration":"GPT-6 Luna Max Effort · gpt-6-luna-max","value":51.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-luna"]},{"id":"claude-sonnet-5-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5-5","label":"Claude Sonnet 5.5 xHigh Effort","configuration":"Claude Sonnet 5.5 xHigh Effort · claude-sonnet-5-5-xhigh-effort","value":39.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/sonnet-5-5/overview"]}]},{"id":"livebench-math","name":"LiveBench · 数学","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:78797b5914d0063bd3a43a98ce30b736dc3ffa4a6f11d850ce49caeec0055bef:1021e4ef211ccba609d6da2c22032f46d695c1546508cf144a58a2f3d6960f3d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-06-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 54 个型号 / 官方材料 66 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":88.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":88.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":78.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":90.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":94.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":95.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":84.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"grok-4.5","modelId":"model:xai/grok-4.5","label":"Grok 4.5","configuration":"Grok 4.5 · grok-4.5","value":90.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.5"]},{"id":"grok-build-0.1","modelId":"model:xai/grok-build-0.1","label":"Grok Build 0.1","configuration":"Grok Build 0.1 · grok-build-0.1","value":78.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-build-0.1"]},{"id":"kimi-k2.7-code","modelId":"model:moonshotai/kimi-k2.7-code","label":"Kimi K2.7 Code","configuration":"Kimi K2.7 Code · kimi-k2.7-code","value":79.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://moonshot.ai"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":79.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":83.72,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":85.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"kimi-k3","modelId":"model:moonshotai/kimi-k3","label":"Kimi K3","configuration":"Kimi K3 · kimi-k3","value":84.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k3"]},{"id":"deepseek-v4-flash-0731","modelId":"model:deepseek/deepseek-v4-flash-0731","label":"DeepSeek V4 Flash 0731","configuration":"DeepSeek V4 Flash 0731 · deepseek-v4-flash-0731","value":86.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/quick_start/pricing/"]},{"id":"qwen3.8-max","modelId":"model:alibaba/qwen3.8-max","label":"Qwen 3.8 Max","configuration":"Qwen 3.8 Max · qwen3.8-max","value":91.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"deepseek-v4-pro-0813","modelId":"model:deepseek/deepseek-v4-pro-0813","label":"DeepSeek V4 Pro 0813","configuration":"DeepSeek V4 Pro 0813 · deepseek-v4-pro-0813","value":95.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/news/news260424"]},{"id":"grok-4.6","modelId":"model:xai/grok-4.6","label":"Grok 4.6 xHigh","configuration":"Grok 4.6 xHigh · grok-4.6","value":92.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://x.ai/news/grok-4-6"]},{"id":"qwen3.8-27b","modelId":"model:alibaba/qwen3.8-27b","label":"Qwen3.8 27B","configuration":"Qwen3.8 27B · qwen3.8-27b","value":86.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.qwencloud.com/models/qwen3.8-27b"]},{"id":"deepseek-v4-flash-vision-exp","modelId":"model:deepseek/deepseek-v4-flash-vision-exp","label":"DeepSeek V4 Flash Vision Exp","configuration":"DeepSeek V4 Flash Vision Exp · deepseek-v4-flash-vision-exp","value":87.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"qwen3.8-flash-next","modelId":"model:alibaba/qwen3.8-flash-next","label":"Qwen 3.8 Flash Next","configuration":"Qwen 3.8 Flash Next · qwen3.8-flash-next","value":85.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":88.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":90.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":89.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":92.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":86.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-sonnet-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5","label":"Claude Sonnet 5 xHigh Effort","configuration":"Claude Sonnet 5 xHigh Effort · claude-sonnet-5-xhigh-effort","value":92.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-5"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":91.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"glm-5.2","modelId":"model:zhipuai/glm-5.2","label":"GLM-5.2","configuration":"GLM-5.2 · glm-5.2","value":89.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":84.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"gpt-5.6-sol-max","modelId":"model:openai/gpt-5.6-sol","label":"GPT-5.6 Sol Max Effort","configuration":"GPT-5.6 Sol Max Effort · gpt-5.6-sol-max","value":96.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-terra-max","modelId":"model:openai/gpt-5.6-terra","label":"GPT-5.6 Terra Max Effort","configuration":"GPT-5.6 Terra Max Effort · gpt-5.6-terra-max","value":94.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-luna-max","modelId":"model:openai/gpt-5.6-luna","label":"GPT-5.6 Luna Max Effort","configuration":"GPT-5.6 Luna Max Effort · gpt-5.6-luna-max","value":87.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"claude-fable-5-max-effort","modelId":"model:anthropic/claude-fable-5","label":"Claude Fable 5 Max Effort","configuration":"Claude Fable 5 Max Effort · claude-fable-5-max-effort","value":95.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5"]},{"id":"muse-spark-1.1-xhigh","modelId":"model:meta/muse-spark-1.1","label":"Muse Spark 1.1 xHigh Effort","configuration":"Muse Spark 1.1 xHigh Effort · muse-spark-1.1-xhigh","value":87.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"inkling-xhigh","modelId":"model:thinkingmachines/inkling","label":"Inkling xHigh Effort","configuration":"Inkling xHigh Effort · inkling-xhigh","value":88.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://thinkingmachines.ai/"]},{"id":"gemini-3.5-flash-lite-high","modelId":"model:google/gemini-3.5-flash-lite","label":"Gemini 3.5 Flash-Lite High","configuration":"Gemini 3.5 Flash-Lite High · gemini-3.5-flash-lite-high","value":73.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite"]},{"id":"gemini-3.6-flash-high","modelId":"model:google/gemini-3.6-flash","label":"Gemini 3.6 Flash High","configuration":"Gemini 3.6 Flash High · gemini-3.6-flash-high","value":86.4,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"]},{"id":"claude-opus-5-max-effort","modelId":"model:anthropic/claude-opus-5","label":"Claude 5 Opus Thinking Max Effort","configuration":"Claude 5 Opus Thinking Max Effort · claude-opus-5-max-effort","value":95.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-5"]},{"id":"claude-opus-4-8-max-effort","modelId":"model:anthropic/claude-opus-4-8","label":"Claude 4.8 Opus Thinking Max Effort","configuration":"Claude 4.8 Opus Thinking Max Effort · claude-opus-4-8-max-effort","value":94.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-8"]},{"id":"muse-spark-1.2-xhigh","modelId":"model:meta/muse-spark-1.2","label":"Muse Spark 1.2 xHigh Effort","configuration":"Muse Spark 1.2 xHigh Effort · muse-spark-1.2-xhigh","value":91.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"gemini-3.7-flash-high","modelId":"model:google/gemini-3.7-flash","label":"Gemini 3.7 Flash High","configuration":"Gemini 3.7 Flash High · gemini-3.7-flash-high","value":93.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash"]},{"id":"glm-5.3","modelId":"model:zhipuai/glm-5.3","label":"GLM-5.3","configuration":"GLM-5.3 · glm-5.3","value":87.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"glm-5.3-flash","modelId":"model:zhipuai/glm-5.3-flash","label":"GLM-5.3 Flash","configuration":"GLM-5.3 Flash · glm-5.3-flash","value":81.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"claude-fable-5-1-max-effort","modelId":"model:anthropic/claude-fable-5-1","label":"Claude Fable 5.1 Max Effort","configuration":"Claude Fable 5.1 Max Effort · claude-fable-5-1-max-effort","value":97.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5-1"]},{"id":"gemini-3.8-flash-high","modelId":"model:google/gemini-3.8-flash","label":"Gemini 3.8 Flash High","configuration":"Gemini 3.8 Flash High · gemini-3.8-flash-high","value":91.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash"]},{"id":"muse-spark-1.3-xhigh","modelId":"model:meta/muse-spark-1.3","label":"Muse Spark 1.3 xHigh Effort","configuration":"Muse Spark 1.3 xHigh Effort · muse-spark-1.3-xhigh","value":95.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://dev.meta.ai/docs/models"]},{"id":"gpt-6-astra-max","modelId":"model:openai/gpt-6-astra","label":"GPT-6 Astra Max Effort","configuration":"GPT-6 Astra Max Effort · gpt-6-astra-max","value":96.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-astra"]},{"id":"deepseek-v4.1-flash-max","modelId":"model:deepseek/deepseek-v4.1-flash","label":"DeepSeek V4.1 Flash Max Effort","configuration":"DeepSeek V4.1 Flash Max Effort · deepseek-v4.1-flash-max","value":93.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"grok-4.7-xhigh","modelId":"model:xai/grok-4.7","label":"Grok 4.7 xHigh","configuration":"Grok 4.7 xHigh · grok-4.7-xhigh","value":95.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.7"]},{"id":"claude-opus-5-5-max-effort","modelId":"model:anthropic/claude-opus-5-5","label":"Claude 5.5 Opus Thinking Max Effort","configuration":"Claude 5.5 Opus Thinking Max Effort · claude-opus-5-5-max-effort","value":97.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/opus-5-5/overview"]},{"id":"gpt-6-sol-max","modelId":"model:openai/gpt-6-sol","label":"GPT-6 Sol Max Effort","configuration":"GPT-6 Sol Max Effort · gpt-6-sol-max","value":96.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-sol"]},{"id":"gpt-6-luna-max","modelId":"model:openai/gpt-6-luna","label":"GPT-6 Luna Max Effort","configuration":"GPT-6 Luna Max Effort · gpt-6-luna-max","value":89.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-luna"]},{"id":"claude-sonnet-5-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5-5","label":"Claude Sonnet 5.5 xHigh Effort","configuration":"Claude Sonnet 5.5 xHigh Effort · claude-sonnet-5-5-xhigh-effort","value":96.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/sonnet-5-5/overview"]}]},{"id":"livebench-data","name":"LiveBench · 数据分析","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:78797b5914d0063bd3a43a98ce30b736dc3ffa4a6f11d850ce49caeec0055bef:1021e4ef211ccba609d6da2c22032f46d695c1546508cf144a58a2f3d6960f3d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-06-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 54 个型号 / 官方材料 66 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":64.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":78.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":70.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":67.64,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":79.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":81.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":55.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"grok-4.5","modelId":"model:xai/grok-4.5","label":"Grok 4.5","configuration":"Grok 4.5 · grok-4.5","value":73.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.5"]},{"id":"grok-build-0.1","modelId":"model:xai/grok-build-0.1","label":"Grok Build 0.1","configuration":"Grok Build 0.1 · grok-build-0.1","value":70.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-build-0.1"]},{"id":"kimi-k2.7-code","modelId":"model:moonshotai/kimi-k2.7-code","label":"Kimi K2.7 Code","configuration":"Kimi K2.7 Code · kimi-k2.7-code","value":62.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://moonshot.ai"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":70.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":69.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":71.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"kimi-k3","modelId":"model:moonshotai/kimi-k3","label":"Kimi K3","configuration":"Kimi K3 · kimi-k3","value":78.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k3"]},{"id":"deepseek-v4-flash-0731","modelId":"model:deepseek/deepseek-v4-flash-0731","label":"DeepSeek V4 Flash 0731","configuration":"DeepSeek V4 Flash 0731 · deepseek-v4-flash-0731","value":79.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/quick_start/pricing/"]},{"id":"qwen3.8-max","modelId":"model:alibaba/qwen3.8-max","label":"Qwen 3.8 Max","configuration":"Qwen 3.8 Max · qwen3.8-max","value":78.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"deepseek-v4-pro-0813","modelId":"model:deepseek/deepseek-v4-pro-0813","label":"DeepSeek V4 Pro 0813","configuration":"DeepSeek V4 Pro 0813 · deepseek-v4-pro-0813","value":79.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/news/news260424"]},{"id":"grok-4.6","modelId":"model:xai/grok-4.6","label":"Grok 4.6 xHigh","configuration":"Grok 4.6 xHigh · grok-4.6","value":73.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://x.ai/news/grok-4-6"]},{"id":"qwen3.8-27b","modelId":"model:alibaba/qwen3.8-27b","label":"Qwen3.8 27B","configuration":"Qwen3.8 27B · qwen3.8-27b","value":76.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.qwencloud.com/models/qwen3.8-27b"]},{"id":"deepseek-v4-flash-vision-exp","modelId":"model:deepseek/deepseek-v4-flash-vision-exp","label":"DeepSeek V4 Flash Vision Exp","configuration":"DeepSeek V4 Flash Vision Exp · deepseek-v4-flash-vision-exp","value":79.48,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"qwen3.8-flash-next","modelId":"model:alibaba/qwen3.8-flash-next","label":"Qwen 3.8 Flash Next","configuration":"Qwen 3.8 Flash Next · qwen3.8-flash-next","value":74.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":54.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":74.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":69.89,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":78.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":77.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-sonnet-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5","label":"Claude Sonnet 5 xHigh Effort","configuration":"Claude Sonnet 5 xHigh Effort · claude-sonnet-5-xhigh-effort","value":71.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-5"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":78.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"glm-5.2","modelId":"model:zhipuai/glm-5.2","label":"GLM-5.2","configuration":"GLM-5.2 · glm-5.2","value":73.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":65.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"gpt-5.6-sol-max","modelId":"model:openai/gpt-5.6-sol","label":"GPT-5.6 Sol Max Effort","configuration":"GPT-5.6 Sol Max Effort · gpt-5.6-sol-max","value":79.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-terra-max","modelId":"model:openai/gpt-5.6-terra","label":"GPT-5.6 Terra Max Effort","configuration":"GPT-5.6 Terra Max Effort · gpt-5.6-terra-max","value":79.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-luna-max","modelId":"model:openai/gpt-5.6-luna","label":"GPT-5.6 Luna Max Effort","configuration":"GPT-5.6 Luna Max Effort · gpt-5.6-luna-max","value":78.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"claude-fable-5-max-effort","modelId":"model:anthropic/claude-fable-5","label":"Claude Fable 5 Max Effort","configuration":"Claude Fable 5 Max Effort · claude-fable-5-max-effort","value":80.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5"]},{"id":"muse-spark-1.1-xhigh","modelId":"model:meta/muse-spark-1.1","label":"Muse Spark 1.1 xHigh Effort","configuration":"Muse Spark 1.1 xHigh Effort · muse-spark-1.1-xhigh","value":72.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"inkling-xhigh","modelId":"model:thinkingmachines/inkling","label":"Inkling xHigh Effort","configuration":"Inkling xHigh Effort · inkling-xhigh","value":72.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://thinkingmachines.ai/"]},{"id":"gemini-3.5-flash-lite-high","modelId":"model:google/gemini-3.5-flash-lite","label":"Gemini 3.5 Flash-Lite High","configuration":"Gemini 3.5 Flash-Lite High · gemini-3.5-flash-lite-high","value":53.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite"]},{"id":"gemini-3.6-flash-high","modelId":"model:google/gemini-3.6-flash","label":"Gemini 3.6 Flash High","configuration":"Gemini 3.6 Flash High · gemini-3.6-flash-high","value":63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"]},{"id":"claude-opus-5-max-effort","modelId":"model:anthropic/claude-opus-5","label":"Claude 5 Opus Thinking Max Effort","configuration":"Claude 5 Opus Thinking Max Effort · claude-opus-5-max-effort","value":74.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-5"]},{"id":"claude-opus-4-8-max-effort","modelId":"model:anthropic/claude-opus-4-8","label":"Claude 4.8 Opus Thinking Max Effort","configuration":"Claude 4.8 Opus Thinking Max Effort · claude-opus-4-8-max-effort","value":66.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-8"]},{"id":"muse-spark-1.2-xhigh","modelId":"model:meta/muse-spark-1.2","label":"Muse Spark 1.2 xHigh Effort","configuration":"Muse Spark 1.2 xHigh Effort · muse-spark-1.2-xhigh","value":76.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"gemini-3.7-flash-high","modelId":"model:google/gemini-3.7-flash","label":"Gemini 3.7 Flash High","configuration":"Gemini 3.7 Flash High · gemini-3.7-flash-high","value":67.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash"]},{"id":"glm-5.3","modelId":"model:zhipuai/glm-5.3","label":"GLM-5.3","configuration":"GLM-5.3 · glm-5.3","value":70.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"glm-5.3-flash","modelId":"model:zhipuai/glm-5.3-flash","label":"GLM-5.3 Flash","configuration":"GLM-5.3 Flash · glm-5.3-flash","value":76.4,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"claude-fable-5-1-max-effort","modelId":"model:anthropic/claude-fable-5-1","label":"Claude Fable 5.1 Max Effort","configuration":"Claude Fable 5.1 Max Effort · claude-fable-5-1-max-effort","value":80.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5-1"]},{"id":"gemini-3.8-flash-high","modelId":"model:google/gemini-3.8-flash","label":"Gemini 3.8 Flash High","configuration":"Gemini 3.8 Flash High · gemini-3.8-flash-high","value":54.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash"]},{"id":"muse-spark-1.3-xhigh","modelId":"model:meta/muse-spark-1.3","label":"Muse Spark 1.3 xHigh Effort","configuration":"Muse Spark 1.3 xHigh Effort · muse-spark-1.3-xhigh","value":79.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://dev.meta.ai/docs/models"]},{"id":"gpt-6-astra-max","modelId":"model:openai/gpt-6-astra","label":"GPT-6 Astra Max Effort","configuration":"GPT-6 Astra Max Effort · gpt-6-astra-max","value":82.97,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-astra"]},{"id":"deepseek-v4.1-flash-max","modelId":"model:deepseek/deepseek-v4.1-flash","label":"DeepSeek V4.1 Flash Max Effort","configuration":"DeepSeek V4.1 Flash Max Effort · deepseek-v4.1-flash-max","value":79.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"grok-4.7-xhigh","modelId":"model:xai/grok-4.7","label":"Grok 4.7 xHigh","configuration":"Grok 4.7 xHigh · grok-4.7-xhigh","value":76.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.7"]},{"id":"claude-opus-5-5-max-effort","modelId":"model:anthropic/claude-opus-5-5","label":"Claude 5.5 Opus Thinking Max Effort","configuration":"Claude 5.5 Opus Thinking Max Effort · claude-opus-5-5-max-effort","value":80.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/opus-5-5/overview"]},{"id":"gpt-6-sol-max","modelId":"model:openai/gpt-6-sol","label":"GPT-6 Sol Max Effort","configuration":"GPT-6 Sol Max Effort · gpt-6-sol-max","value":81.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-sol"]},{"id":"gpt-6-luna-max","modelId":"model:openai/gpt-6-luna","label":"GPT-6 Luna Max Effort","configuration":"GPT-6 Luna Max Effort · gpt-6-luna-max","value":73.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-luna"]},{"id":"claude-sonnet-5-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5-5","label":"Claude Sonnet 5.5 xHigh Effort","configuration":"Claude Sonnet 5.5 xHigh Effort · claude-sonnet-5-5-xhigh-effort","value":78.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/sonnet-5-5/overview"]}]},{"id":"livebench-language","name":"LiveBench · 语言","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:78797b5914d0063bd3a43a98ce30b736dc3ffa4a6f11d850ce49caeec0055bef:1021e4ef211ccba609d6da2c22032f46d695c1546508cf144a58a2f3d6960f3d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-06-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 54 个型号 / 官方材料 66 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":84.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":73.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":70.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":62.51,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":82.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":87.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":73.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"grok-4.5","modelId":"model:xai/grok-4.5","label":"Grok 4.5","configuration":"Grok 4.5 · grok-4.5","value":82.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.5"]},{"id":"grok-build-0.1","modelId":"model:xai/grok-build-0.1","label":"Grok Build 0.1","configuration":"Grok Build 0.1 · grok-build-0.1","value":72.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-build-0.1"]},{"id":"kimi-k2.7-code","modelId":"model:moonshotai/kimi-k2.7-code","label":"Kimi K2.7 Code","configuration":"Kimi K2.7 Code · kimi-k2.7-code","value":77.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://moonshot.ai"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":63.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":74.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":79.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"kimi-k3","modelId":"model:moonshotai/kimi-k3","label":"Kimi K3","configuration":"Kimi K3 · kimi-k3","value":85.53,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k3"]},{"id":"deepseek-v4-flash-0731","modelId":"model:deepseek/deepseek-v4-flash-0731","label":"DeepSeek V4 Flash 0731","configuration":"DeepSeek V4 Flash 0731 · deepseek-v4-flash-0731","value":79.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/quick_start/pricing/"]},{"id":"qwen3.8-max","modelId":"model:alibaba/qwen3.8-max","label":"Qwen 3.8 Max","configuration":"Qwen 3.8 Max · qwen3.8-max","value":79.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"deepseek-v4-pro-0813","modelId":"model:deepseek/deepseek-v4-pro-0813","label":"DeepSeek V4 Pro 0813","configuration":"DeepSeek V4 Pro 0813 · deepseek-v4-pro-0813","value":82.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/news/news260424"]},{"id":"grok-4.6","modelId":"model:xai/grok-4.6","label":"Grok 4.6 xHigh","configuration":"Grok 4.6 xHigh · grok-4.6","value":83.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://x.ai/news/grok-4-6"]},{"id":"qwen3.8-27b","modelId":"model:alibaba/qwen3.8-27b","label":"Qwen3.8 27B","configuration":"Qwen3.8 27B · qwen3.8-27b","value":74.35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.qwencloud.com/models/qwen3.8-27b"]},{"id":"deepseek-v4-flash-vision-exp","modelId":"model:deepseek/deepseek-v4-flash-vision-exp","label":"DeepSeek V4 Flash Vision Exp","configuration":"DeepSeek V4 Flash Vision Exp · deepseek-v4-flash-vision-exp","value":80.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"qwen3.8-flash-next","modelId":"model:alibaba/qwen3.8-flash-next","label":"Qwen 3.8 Flash Next","configuration":"Qwen 3.8 Flash Next · qwen3.8-flash-next","value":74.64,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":70.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":81.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":83.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":77.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":76.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-sonnet-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5","label":"Claude Sonnet 5 xHigh Effort","configuration":"Claude Sonnet 5 xHigh Effort · claude-sonnet-5-xhigh-effort","value":74.97,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-5"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":85.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"glm-5.2","modelId":"model:zhipuai/glm-5.2","label":"GLM-5.2","configuration":"GLM-5.2 · glm-5.2","value":76.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":75.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"gpt-5.6-sol-max","modelId":"model:openai/gpt-5.6-sol","label":"GPT-5.6 Sol Max Effort","configuration":"GPT-5.6 Sol Max Effort · gpt-5.6-sol-max","value":87.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-terra-max","modelId":"model:openai/gpt-5.6-terra","label":"GPT-5.6 Terra Max Effort","configuration":"GPT-5.6 Terra Max Effort · gpt-5.6-terra-max","value":82.89,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-luna-max","modelId":"model:openai/gpt-5.6-luna","label":"GPT-5.6 Luna Max Effort","configuration":"GPT-5.6 Luna Max Effort · gpt-5.6-luna-max","value":72.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"claude-fable-5-max-effort","modelId":"model:anthropic/claude-fable-5","label":"Claude Fable 5 Max Effort","configuration":"Claude Fable 5 Max Effort · claude-fable-5-max-effort","value":90.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5"]},{"id":"muse-spark-1.1-xhigh","modelId":"model:meta/muse-spark-1.1","label":"Muse Spark 1.1 xHigh Effort","configuration":"Muse Spark 1.1 xHigh Effort · muse-spark-1.1-xhigh","value":74.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"inkling-xhigh","modelId":"model:thinkingmachines/inkling","label":"Inkling xHigh Effort","configuration":"Inkling xHigh Effort · inkling-xhigh","value":73.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://thinkingmachines.ai/"]},{"id":"gemini-3.5-flash-lite-high","modelId":"model:google/gemini-3.5-flash-lite","label":"Gemini 3.5 Flash-Lite High","configuration":"Gemini 3.5 Flash-Lite High · gemini-3.5-flash-lite-high","value":71.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite"]},{"id":"gemini-3.6-flash-high","modelId":"model:google/gemini-3.6-flash","label":"Gemini 3.6 Flash High","configuration":"Gemini 3.6 Flash High · gemini-3.6-flash-high","value":83.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"]},{"id":"claude-opus-5-max-effort","modelId":"model:anthropic/claude-opus-5","label":"Claude 5 Opus Thinking Max Effort","configuration":"Claude 5 Opus Thinking Max Effort · claude-opus-5-max-effort","value":88.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-5"]},{"id":"claude-opus-4-8-max-effort","modelId":"model:anthropic/claude-opus-4-8","label":"Claude 4.8 Opus Thinking Max Effort","configuration":"Claude 4.8 Opus Thinking Max Effort · claude-opus-4-8-max-effort","value":79.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-8"]},{"id":"muse-spark-1.2-xhigh","modelId":"model:meta/muse-spark-1.2","label":"Muse Spark 1.2 xHigh Effort","configuration":"Muse Spark 1.2 xHigh Effort · muse-spark-1.2-xhigh","value":78.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"gemini-3.7-flash-high","modelId":"model:google/gemini-3.7-flash","label":"Gemini 3.7 Flash High","configuration":"Gemini 3.7 Flash High · gemini-3.7-flash-high","value":85.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash"]},{"id":"glm-5.3","modelId":"model:zhipuai/glm-5.3","label":"GLM-5.3","configuration":"GLM-5.3 · glm-5.3","value":79.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"glm-5.3-flash","modelId":"model:zhipuai/glm-5.3-flash","label":"GLM-5.3 Flash","configuration":"GLM-5.3 Flash · glm-5.3-flash","value":77.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"claude-fable-5-1-max-effort","modelId":"model:anthropic/claude-fable-5-1","label":"Claude Fable 5.1 Max Effort","configuration":"Claude Fable 5.1 Max Effort · claude-fable-5-1-max-effort","value":89.5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5-1"]},{"id":"gemini-3.8-flash-high","modelId":"model:google/gemini-3.8-flash","label":"Gemini 3.8 Flash High","configuration":"Gemini 3.8 Flash High · gemini-3.8-flash-high","value":87.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash"]},{"id":"muse-spark-1.3-xhigh","modelId":"model:meta/muse-spark-1.3","label":"Muse Spark 1.3 xHigh Effort","configuration":"Muse Spark 1.3 xHigh Effort · muse-spark-1.3-xhigh","value":82.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://dev.meta.ai/docs/models"]},{"id":"gpt-6-astra-max","modelId":"model:openai/gpt-6-astra","label":"GPT-6 Astra Max Effort","configuration":"GPT-6 Astra Max Effort · gpt-6-astra-max","value":89.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-astra"]},{"id":"deepseek-v4.1-flash-max","modelId":"model:deepseek/deepseek-v4.1-flash","label":"DeepSeek V4.1 Flash Max Effort","configuration":"DeepSeek V4.1 Flash Max Effort · deepseek-v4.1-flash-max","value":81.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"grok-4.7-xhigh","modelId":"model:xai/grok-4.7","label":"Grok 4.7 xHigh","configuration":"Grok 4.7 xHigh · grok-4.7-xhigh","value":80.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.7"]},{"id":"claude-opus-5-5-max-effort","modelId":"model:anthropic/claude-opus-5-5","label":"Claude 5.5 Opus Thinking Max Effort","configuration":"Claude 5.5 Opus Thinking Max Effort · claude-opus-5-5-max-effort","value":86.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/opus-5-5/overview"]},{"id":"gpt-6-sol-max","modelId":"model:openai/gpt-6-sol","label":"GPT-6 Sol Max Effort","configuration":"GPT-6 Sol Max Effort · gpt-6-sol-max","value":85.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-sol"]},{"id":"gpt-6-luna-max","modelId":"model:openai/gpt-6-luna","label":"GPT-6 Luna Max Effort","configuration":"GPT-6 Luna Max Effort · gpt-6-luna-max","value":73.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-luna"]},{"id":"claude-sonnet-5-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5-5","label":"Claude Sonnet 5.5 xHigh Effort","configuration":"Claude Sonnet 5.5 xHigh Effort · claude-sonnet-5-5-xhigh-effort","value":83.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/sonnet-5-5/overview"]}]},{"id":"livebench-if","name":"LiveBench · 指令遵循","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:78797b5914d0063bd3a43a98ce30b736dc3ffa4a6f11d850ce49caeec0055bef:1021e4ef211ccba609d6da2c22032f46d695c1546508cf144a58a2f3d6960f3d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-06-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 54 个型号 / 官方材料 66 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":75.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":66.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":59.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":67.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":70.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":70.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":62.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"grok-4.5","modelId":"model:xai/grok-4.5","label":"Grok 4.5","configuration":"Grok 4.5 · grok-4.5","value":71.53,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.5"]},{"id":"grok-build-0.1","modelId":"model:xai/grok-build-0.1","label":"Grok Build 0.1","configuration":"Grok Build 0.1 · grok-build-0.1","value":65.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-build-0.1"]},{"id":"kimi-k2.7-code","modelId":"model:moonshotai/kimi-k2.7-code","label":"Kimi K2.7 Code","configuration":"Kimi K2.7 Code · kimi-k2.7-code","value":56.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://moonshot.ai"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":53.23,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":58.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":74.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"kimi-k3","modelId":"model:moonshotai/kimi-k3","label":"Kimi K3","configuration":"Kimi K3 · kimi-k3","value":71.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k3"]},{"id":"deepseek-v4-flash-0731","modelId":"model:deepseek/deepseek-v4-flash-0731","label":"DeepSeek V4 Flash 0731","configuration":"DeepSeek V4 Flash 0731 · deepseek-v4-flash-0731","value":65.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/quick_start/pricing/"]},{"id":"qwen3.8-max","modelId":"model:alibaba/qwen3.8-max","label":"Qwen 3.8 Max","configuration":"Qwen 3.8 Max · qwen3.8-max","value":74.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"deepseek-v4-pro-0813","modelId":"model:deepseek/deepseek-v4-pro-0813","label":"DeepSeek V4 Pro 0813","configuration":"DeepSeek V4 Pro 0813 · deepseek-v4-pro-0813","value":67.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/news/news260424"]},{"id":"grok-4.6","modelId":"model:xai/grok-4.6","label":"Grok 4.6 xHigh","configuration":"Grok 4.6 xHigh · grok-4.6","value":71.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://x.ai/news/grok-4-6"]},{"id":"qwen3.8-27b","modelId":"model:alibaba/qwen3.8-27b","label":"Qwen3.8 27B","configuration":"Qwen3.8 27B · qwen3.8-27b","value":72.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.qwencloud.com/models/qwen3.8-27b"]},{"id":"deepseek-v4-flash-vision-exp","modelId":"model:deepseek/deepseek-v4-flash-vision-exp","label":"DeepSeek V4 Flash Vision Exp","configuration":"DeepSeek V4 Flash Vision Exp · deepseek-v4-flash-vision-exp","value":70.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"qwen3.8-flash-next","modelId":"model:alibaba/qwen3.8-flash-next","label":"Qwen 3.8 Flash Next","configuration":"Qwen 3.8 Flash Next · qwen3.8-flash-next","value":77.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://qwen.ai/blog?id=qwen3.8"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":73.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":62.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":63.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":66.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":63.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-sonnet-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5","label":"Claude Sonnet 5 xHigh Effort","configuration":"Claude Sonnet 5 xHigh Effort · claude-sonnet-5-xhigh-effort","value":63.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-sonnet-5"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":79.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"glm-5.2","modelId":"model:zhipuai/glm-5.2","label":"GLM-5.2","configuration":"GLM-5.2 · glm-5.2","value":62.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":64.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"gpt-5.6-sol-max","modelId":"model:openai/gpt-5.6-sol","label":"GPT-5.6 Sol Max Effort","configuration":"GPT-5.6 Sol Max Effort · gpt-5.6-sol-max","value":71.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-terra-max","modelId":"model:openai/gpt-5.6-terra","label":"GPT-5.6 Terra Max Effort","configuration":"GPT-5.6 Terra Max Effort · gpt-5.6-terra-max","value":64.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"gpt-5.6-luna-max","modelId":"model:openai/gpt-5.6-luna","label":"GPT-5.6 Luna Max Effort","configuration":"GPT-5.6 Luna Max Effort · gpt-5.6-luna-max","value":60.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.openai.com/docs/models/gpt-5.6"]},{"id":"claude-fable-5-max-effort","modelId":"model:anthropic/claude-fable-5","label":"Claude Fable 5 Max Effort","configuration":"Claude Fable 5 Max Effort · claude-fable-5-max-effort","value":75.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5"]},{"id":"muse-spark-1.1-xhigh","modelId":"model:meta/muse-spark-1.1","label":"Muse Spark 1.1 xHigh Effort","configuration":"Muse Spark 1.1 xHigh Effort · muse-spark-1.1-xhigh","value":69.64,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"inkling-xhigh","modelId":"model:thinkingmachines/inkling","label":"Inkling xHigh Effort","configuration":"Inkling xHigh Effort · inkling-xhigh","value":70.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://thinkingmachines.ai/"]},{"id":"gemini-3.5-flash-lite-high","modelId":"model:google/gemini-3.5-flash-lite","label":"Gemini 3.5 Flash-Lite High","configuration":"Gemini 3.5 Flash-Lite High · gemini-3.5-flash-lite-high","value":67.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.5-flash-lite"]},{"id":"gemini-3.6-flash-high","modelId":"model:google/gemini-3.6-flash","label":"Gemini 3.6 Flash High","configuration":"Gemini 3.6 Flash High · gemini-3.6-flash-high","value":75.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.6-flash"]},{"id":"claude-opus-5-max-effort","modelId":"model:anthropic/claude-opus-5","label":"Claude 5 Opus Thinking Max Effort","configuration":"Claude 5 Opus Thinking Max Effort · claude-opus-5-max-effort","value":63.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-5"]},{"id":"claude-opus-4-8-max-effort","modelId":"model:anthropic/claude-opus-4-8","label":"Claude 4.8 Opus Thinking Max Effort","configuration":"Claude 4.8 Opus Thinking Max Effort · claude-opus-4-8-max-effort","value":72.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-opus-4-8"]},{"id":"muse-spark-1.2-xhigh","modelId":"model:meta/muse-spark-1.2","label":"Muse Spark 1.2 xHigh Effort","configuration":"Muse Spark 1.2 xHigh Effort · muse-spark-1.2-xhigh","value":74.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.meta.com/"]},{"id":"gemini-3.7-flash-high","modelId":"model:google/gemini-3.7-flash","label":"Gemini 3.7 Flash High","configuration":"Gemini 3.7 Flash High · gemini-3.7-flash-high","value":79.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.7-flash"]},{"id":"glm-5.3","modelId":"model:zhipuai/glm-5.3","label":"GLM-5.3","configuration":"GLM-5.3 · glm-5.3","value":69.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"glm-5.3-flash","modelId":"model:zhipuai/glm-5.3-flash","label":"GLM-5.3 Flash","configuration":"GLM-5.3 Flash · glm-5.3-flash","value":52.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://z.ai"]},{"id":"claude-fable-5-1-max-effort","modelId":"model:anthropic/claude-fable-5-1","label":"Claude Fable 5.1 Max Effort","configuration":"Claude Fable 5.1 Max Effort · claude-fable-5-1-max-effort","value":72.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://www.anthropic.com/news/claude-fable-5-1"]},{"id":"gemini-3.8-flash-high","modelId":"model:google/gemini-3.8-flash","label":"Gemini 3.8 Flash High","configuration":"Gemini 3.8 Flash High · gemini-3.8-flash-high","value":81.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash"]},{"id":"muse-spark-1.3-xhigh","modelId":"model:meta/muse-spark-1.3","label":"Muse Spark 1.3 xHigh Effort","configuration":"Muse Spark 1.3 xHigh Effort · muse-spark-1.3-xhigh","value":78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://dev.meta.ai/docs/models"]},{"id":"gpt-6-astra-max","modelId":"model:openai/gpt-6-astra","label":"GPT-6 Astra Max Effort","configuration":"GPT-6 Astra Max Effort · gpt-6-astra-max","value":75.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-astra"]},{"id":"deepseek-v4.1-flash-max","modelId":"model:deepseek/deepseek-v4.1-flash","label":"DeepSeek V4.1 Flash Max Effort","configuration":"DeepSeek V4.1 Flash Max Effort · deepseek-v4.1-flash-max","value":70.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://api-docs.deepseek.com/updates"]},{"id":"grok-4.7-xhigh","modelId":"model:xai/grok-4.7","label":"Grok 4.7 xHigh","configuration":"Grok 4.7 xHigh · grok-4.7-xhigh","value":75.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://docs.x.ai/developers/models/grok-4.7"]},{"id":"claude-opus-5-5-max-effort","modelId":"model:anthropic/claude-opus-5-5","label":"Claude 5.5 Opus Thinking Max Effort","configuration":"Claude 5.5 Opus Thinking Max Effort · claude-opus-5-5-max-effort","value":65.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/opus-5-5/overview"]},{"id":"gpt-6-sol-max","modelId":"model:openai/gpt-6-sol","label":"GPT-6 Sol Max Effort","configuration":"GPT-6 Sol Max Effort · gpt-6-sol-max","value":68.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-sol"]},{"id":"gpt-6-luna-max","modelId":"model:openai/gpt-6-luna","label":"GPT-6 Luna Max Effort","configuration":"GPT-6 Luna Max Effort · gpt-6-luna-max","value":55.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://developers.openai.com/api/docs/models/gpt-6-luna"]},{"id":"claude-sonnet-5-5-xhigh-effort","modelId":"model:anthropic/claude-sonnet-5-5","label":"Claude Sonnet 5.5 xHigh Effort","configuration":"Claude Sonnet 5.5 xHigh Effort · claude-sonnet-5-5-xhigh-effort","value":70.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_06_25.csv","https://platform.claude.com/docs/en/models/sonnet-5-5/overview"]}]},{"id":"livebench-overall-2024-11-25","name":"LiveBench · 全科 · 2024-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:61806df0681ac4c65b21e995447d2fcf35020adf0e14a3f6d9d8d68146f4df4f:17ca6b3d122921d24b591d0da52566c5552ed4f4001ae9d83a11b2712325aae4","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2024-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 19 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"amazon.nova-lite-v1:0","modelId":"model:amazon/nova-lite","label":"Nova Lite","configuration":"Nova Lite · amazon.nova-lite-v1:0","value":36.35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-micro-v1:0","modelId":"model:amazon/nova-micro","label":"Nova Micro","configuration":"Nova Micro · amazon.nova-micro-v1:0","value":29.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-pro-v1:0","modelId":"model:amazon/nova-pro","label":"Nova Pro","configuration":"Nova Pro · amazon.nova-pro-v1:0","value":43.53,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"claude-3-5-haiku-20241022","modelId":"model:anthropic/claude-3-5-haiku-20241022","label":"Claude 3.5 Haiku","configuration":"Claude 3.5 Haiku · claude-3-5-haiku-20241022","value":43.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/claude/haiku"]},{"id":"claude-3-5-sonnet-20241022","modelId":"model:anthropic/claude-3-5-sonnet-20241022","label":"Claude 3.5 Sonnet","configuration":"Claude 3.5 Sonnet · claude-3-5-sonnet-20241022","value":59.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/news/3-5-models-and-computer-use"]},{"id":"command-r-plus-08-2024","modelId":"model:cohere/command-r-plus-08-2024","label":"Command R Plus","configuration":"Command R Plus · command-r-plus-08-2024","value":31.76,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://docs.cohere.com/docs/models"]},{"id":"deepseek-r1","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1","value":71.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1"]},{"id":"deepseek-r1-distill-qwen-32b","modelId":"model:deepseek/deepseek-r1-distill-qwen-32b","label":"DeepSeek R1 Distill Qwen 32B","configuration":"DeepSeek R1 Distill Qwen 32B · deepseek-r1-distill-qwen-32b","value":45.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"]},{"id":"deepseek-v3","modelId":"model:deepseek/deepseek-v3","label":"DeepSeek V3","configuration":"DeepSeek V3 · deepseek-v3","value":60.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://api-docs.deepseek.com/news/news1226"]},{"id":"gemini-2.0-flash","modelId":"model:google/gemini-2.0-flash","label":"Gemini 2.0 Flash","configuration":"Gemini 2.0 Flash · gemini-2.0-flash","value":61.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/"]},{"id":"gpt-4o-mini-2024-07-18","modelId":"model:openai/gpt-4o-mini","label":"GPT-4o Mini","configuration":"GPT-4o Mini · gpt-4o-mini-2024-07-18","value":41.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/hello-gpt-4o/"]},{"id":"mistral-large-2411","modelId":"model:mistral/mistral-large-2411","label":"Mistral Large","configuration":"Mistral Large · mistral-large-2411","value":48.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/mistralai/Mistral-Large-Instruct-2411"]},{"id":"o1-2024-12-17-high","modelId":"model:openai/o1","label":"o1 High","configuration":"o1 High · o1-2024-12-17-high","value":75.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/o1/"]},{"id":"o3-mini-2025-01-31-high","modelId":"model:openai/o3-mini","label":"o3-Mini High","configuration":"o3-Mini High · o3-mini-2025-01-31-high","value":75.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/openai-o3-mini/"]},{"id":"qwen2.5-coder-32b-instruct","modelId":"model:alibaba/qwen2.5-coder-32b-instruct","label":"Qwen2.5 Coder 32B Instruct","configuration":"Qwen2.5 Coder 32B Instruct · qwen2.5-coder-32b-instruct","value":46.23,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct"]},{"id":"gemini-2.0-flash-lite","modelId":"model:google/gemini-2.0-flash-lite","label":"Gemini 2.0 Flash Lite","configuration":"Gemini 2.0 Flash Lite · gemini-2.0-flash-lite","value":54.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://developers.googleblog.com/en/start-building-with-the-gemini-2-0-flash-family/"]},{"id":"qwq-32b","modelId":"model:alibaba/qwq-32b","label":"QwQ 32B","configuration":"QwQ 32B · qwq-32b","value":71.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://qwenlm.github.io/blog/qwq-32b/"]},{"id":"gemma-3-27b-it","modelId":"model:google/gemma-3-27b-it","label":"Gemma 3 27B","configuration":"Gemma 3 27B · gemma-3-27b-it","value":49.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/developers/gemma-3/"]},{"id":"deepseek-v3-0324","modelId":"model:deepseek/deepseek-v3-0324","label":"DeepSeek V3 0324","configuration":"DeepSeek V3 0324 · deepseek-v3-0324","value":66.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3-0324"]}]},{"id":"livebench-reasoning-2024-11-25","name":"LiveBench · 推理 · 2024-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:61806df0681ac4c65b21e995447d2fcf35020adf0e14a3f6d9d8d68146f4df4f:17ca6b3d122921d24b591d0da52566c5552ed4f4001ae9d83a11b2712325aae4","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2024-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 19 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"amazon.nova-lite-v1:0","modelId":"model:amazon/nova-lite","label":"Nova Lite","configuration":"Nova Lite · amazon.nova-lite-v1:0","value":36.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-micro-v1:0","modelId":"model:amazon/nova-micro","label":"Nova Micro","configuration":"Nova Micro · amazon.nova-micro-v1:0","value":25.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-pro-v1:0","modelId":"model:amazon/nova-pro","label":"Nova Pro","configuration":"Nova Pro · amazon.nova-pro-v1:0","value":32.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"claude-3-5-haiku-20241022","modelId":"model:anthropic/claude-3-5-haiku-20241022","label":"Claude 3.5 Haiku","configuration":"Claude 3.5 Haiku · claude-3-5-haiku-20241022","value":28.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/claude/haiku"]},{"id":"claude-3-5-sonnet-20241022","modelId":"model:anthropic/claude-3-5-sonnet-20241022","label":"Claude 3.5 Sonnet","configuration":"Claude 3.5 Sonnet · claude-3-5-sonnet-20241022","value":56.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/news/3-5-models-and-computer-use"]},{"id":"command-r-plus-08-2024","modelId":"model:cohere/command-r-plus-08-2024","label":"Command R Plus","configuration":"Command R Plus · command-r-plus-08-2024","value":24.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://docs.cohere.com/docs/models"]},{"id":"deepseek-r1","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1","value":83.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1"]},{"id":"deepseek-r1-distill-qwen-32b","modelId":"model:deepseek/deepseek-r1-distill-qwen-32b","label":"DeepSeek R1 Distill Qwen 32B","configuration":"DeepSeek R1 Distill Qwen 32B · deepseek-r1-distill-qwen-32b","value":52.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"]},{"id":"deepseek-v3","modelId":"model:deepseek/deepseek-v3","label":"DeepSeek V3","configuration":"DeepSeek V3 · deepseek-v3","value":56.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://api-docs.deepseek.com/news/news1226"]},{"id":"gemini-2.0-flash","modelId":"model:google/gemini-2.0-flash","label":"Gemini 2.0 Flash","configuration":"Gemini 2.0 Flash · gemini-2.0-flash","value":55.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/"]},{"id":"gpt-4o-mini-2024-07-18","modelId":"model:openai/gpt-4o-mini","label":"GPT-4o Mini","configuration":"GPT-4o Mini · gpt-4o-mini-2024-07-18","value":32.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/hello-gpt-4o/"]},{"id":"mistral-large-2411","modelId":"model:mistral/mistral-large-2411","label":"Mistral Large","configuration":"Mistral Large · mistral-large-2411","value":43.5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/mistralai/Mistral-Large-Instruct-2411"]},{"id":"o1-2024-12-17-high","modelId":"model:openai/o1","label":"o1 High","configuration":"o1 High · o1-2024-12-17-high","value":91.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/o1/"]},{"id":"o3-mini-2025-01-31-high","modelId":"model:openai/o3-mini","label":"o3-Mini High","configuration":"o3-Mini High · o3-mini-2025-01-31-high","value":89.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/openai-o3-mini/"]},{"id":"qwen2.5-coder-32b-instruct","modelId":"model:alibaba/qwen2.5-coder-32b-instruct","label":"Qwen2.5 Coder 32B Instruct","configuration":"Qwen2.5 Coder 32B Instruct · qwen2.5-coder-32b-instruct","value":42.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct"]},{"id":"gemini-2.0-flash-lite","modelId":"model:google/gemini-2.0-flash-lite","label":"Gemini 2.0 Flash Lite","configuration":"Gemini 2.0 Flash Lite · gemini-2.0-flash-lite","value":44.92,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://developers.googleblog.com/en/start-building-with-the-gemini-2-0-flash-family/"]},{"id":"qwq-32b","modelId":"model:alibaba/qwq-32b","label":"QwQ 32B","configuration":"QwQ 32B · qwq-32b","value":83.5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://qwenlm.github.io/blog/qwq-32b/"]},{"id":"gemma-3-27b-it","modelId":"model:google/gemma-3-27b-it","label":"Gemma 3 27B","configuration":"Gemma 3 27B · gemma-3-27b-it","value":43.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/developers/gemma-3/"]},{"id":"deepseek-v3-0324","modelId":"model:deepseek/deepseek-v3-0324","label":"DeepSeek V3 0324","configuration":"DeepSeek V3 0324 · deepseek-v3-0324","value":65.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3-0324"]}]},{"id":"livebench-coding-2024-11-25","name":"LiveBench · 编程答题 · 2024-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:61806df0681ac4c65b21e995447d2fcf35020adf0e14a3f6d9d8d68146f4df4f:17ca6b3d122921d24b591d0da52566c5552ed4f4001ae9d83a11b2712325aae4","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2024-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 19 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"amazon.nova-lite-v1:0","modelId":"model:amazon/nova-lite","label":"Nova Lite","configuration":"Nova Lite · amazon.nova-lite-v1:0","value":27.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-micro-v1:0","modelId":"model:amazon/nova-micro","label":"Nova Micro","configuration":"Nova Micro · amazon.nova-micro-v1:0","value":20.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-pro-v1:0","modelId":"model:amazon/nova-pro","label":"Nova Pro","configuration":"Nova Pro · amazon.nova-pro-v1:0","value":38.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"claude-3-5-haiku-20241022","modelId":"model:anthropic/claude-3-5-haiku-20241022","label":"Claude 3.5 Haiku","configuration":"Claude 3.5 Haiku · claude-3-5-haiku-20241022","value":51.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/claude/haiku"]},{"id":"claude-3-5-sonnet-20241022","modelId":"model:anthropic/claude-3-5-sonnet-20241022","label":"Claude 3.5 Sonnet","configuration":"Claude 3.5 Sonnet · claude-3-5-sonnet-20241022","value":67.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/news/3-5-models-and-computer-use"]},{"id":"command-r-plus-08-2024","modelId":"model:cohere/command-r-plus-08-2024","label":"Command R Plus","configuration":"Command R Plus · command-r-plus-08-2024","value":19.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://docs.cohere.com/docs/models"]},{"id":"deepseek-r1","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1","value":66.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1"]},{"id":"deepseek-r1-distill-qwen-32b","modelId":"model:deepseek/deepseek-r1-distill-qwen-32b","label":"DeepSeek R1 Distill Qwen 32B","configuration":"DeepSeek R1 Distill Qwen 32B · deepseek-r1-distill-qwen-32b","value":33.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"]},{"id":"deepseek-v3","modelId":"model:deepseek/deepseek-v3","label":"DeepSeek V3","configuration":"DeepSeek V3 · deepseek-v3","value":61.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://api-docs.deepseek.com/news/news1226"]},{"id":"gemini-2.0-flash","modelId":"model:google/gemini-2.0-flash","label":"Gemini 2.0 Flash","configuration":"Gemini 2.0 Flash · gemini-2.0-flash","value":53.92,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/"]},{"id":"gpt-4o-mini-2024-07-18","modelId":"model:openai/gpt-4o-mini","label":"GPT-4o Mini","configuration":"GPT-4o Mini · gpt-4o-mini-2024-07-18","value":43.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/hello-gpt-4o/"]},{"id":"mistral-large-2411","modelId":"model:mistral/mistral-large-2411","label":"Mistral Large","configuration":"Mistral Large · mistral-large-2411","value":47.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/mistralai/Mistral-Large-Instruct-2411"]},{"id":"o1-2024-12-17-high","modelId":"model:openai/o1","label":"o1 High","configuration":"o1 High · o1-2024-12-17-high","value":69.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/o1/"]},{"id":"o3-mini-2025-01-31-high","modelId":"model:openai/o3-mini","label":"o3-Mini High","configuration":"o3-Mini High · o3-mini-2025-01-31-high","value":82.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/openai-o3-mini/"]},{"id":"qwen2.5-coder-32b-instruct","modelId":"model:alibaba/qwen2.5-coder-32b-instruct","label":"Qwen2.5 Coder 32B Instruct","configuration":"Qwen2.5 Coder 32B Instruct · qwen2.5-coder-32b-instruct","value":56.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct"]},{"id":"gemini-2.0-flash-lite","modelId":"model:google/gemini-2.0-flash-lite","label":"Gemini 2.0 Flash Lite","configuration":"Gemini 2.0 Flash Lite · gemini-2.0-flash-lite","value":47.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://developers.googleblog.com/en/start-building-with-the-gemini-2-0-flash-family/"]},{"id":"qwq-32b","modelId":"model:alibaba/qwq-32b","label":"QwQ 32B","configuration":"QwQ 32B · qwq-32b","value":72.23,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://qwenlm.github.io/blog/qwq-32b/"]},{"id":"gemma-3-27b-it","modelId":"model:google/gemma-3-27b-it","label":"Gemma 3 27B","configuration":"Gemma 3 27B · gemma-3-27b-it","value":39.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/developers/gemma-3/"]},{"id":"deepseek-v3-0324","modelId":"model:deepseek/deepseek-v3-0324","label":"DeepSeek V3 0324","configuration":"DeepSeek V3 0324 · deepseek-v3-0324","value":70.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3-0324"]}]},{"id":"livebench-math-2024-11-25","name":"LiveBench · 数学 · 2024-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:61806df0681ac4c65b21e995447d2fcf35020adf0e14a3f6d9d8d68146f4df4f:17ca6b3d122921d24b591d0da52566c5552ed4f4001ae9d83a11b2712325aae4","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2024-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 19 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"amazon.nova-lite-v1:0","modelId":"model:amazon/nova-lite","label":"Nova Lite","configuration":"Nova Lite · amazon.nova-lite-v1:0","value":36.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-micro-v1:0","modelId":"model:amazon/nova-micro","label":"Nova Micro","configuration":"Nova Micro · amazon.nova-micro-v1:0","value":34.49,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-pro-v1:0","modelId":"model:amazon/nova-pro","label":"Nova Pro","configuration":"Nova Pro · amazon.nova-pro-v1:0","value":38.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"claude-3-5-haiku-20241022","modelId":"model:anthropic/claude-3-5-haiku-20241022","label":"Claude 3.5 Haiku","configuration":"Claude 3.5 Haiku · claude-3-5-haiku-20241022","value":35.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/claude/haiku"]},{"id":"claude-3-5-sonnet-20241022","modelId":"model:anthropic/claude-3-5-sonnet-20241022","label":"Claude 3.5 Sonnet","configuration":"Claude 3.5 Sonnet · claude-3-5-sonnet-20241022","value":52.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/news/3-5-models-and-computer-use"]},{"id":"command-r-plus-08-2024","modelId":"model:cohere/command-r-plus-08-2024","label":"Command R Plus","configuration":"Command R Plus · command-r-plus-08-2024","value":21.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://docs.cohere.com/docs/models"]},{"id":"deepseek-r1","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1","value":80.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1"]},{"id":"deepseek-r1-distill-qwen-32b","modelId":"model:deepseek/deepseek-r1-distill-qwen-32b","label":"DeepSeek R1 Distill Qwen 32B","configuration":"DeepSeek R1 Distill Qwen 32B · deepseek-r1-distill-qwen-32b","value":59.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"]},{"id":"deepseek-v3","modelId":"model:deepseek/deepseek-v3","label":"DeepSeek V3","configuration":"DeepSeek V3 · deepseek-v3","value":60.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://api-docs.deepseek.com/news/news1226"]},{"id":"gemini-2.0-flash","modelId":"model:google/gemini-2.0-flash","label":"Gemini 2.0 Flash","configuration":"Gemini 2.0 Flash · gemini-2.0-flash","value":65.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/"]},{"id":"gpt-4o-mini-2024-07-18","modelId":"model:openai/gpt-4o-mini","label":"GPT-4o Mini","configuration":"GPT-4o Mini · gpt-4o-mini-2024-07-18","value":36.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/hello-gpt-4o/"]},{"id":"mistral-large-2411","modelId":"model:mistral/mistral-large-2411","label":"Mistral Large","configuration":"Mistral Large · mistral-large-2411","value":42.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/mistralai/Mistral-Large-Instruct-2411"]},{"id":"o1-2024-12-17-high","modelId":"model:openai/o1","label":"o1 High","configuration":"o1 High · o1-2024-12-17-high","value":80.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/o1/"]},{"id":"o3-mini-2025-01-31-high","modelId":"model:openai/o3-mini","label":"o3-Mini High","configuration":"o3-Mini High · o3-mini-2025-01-31-high","value":77.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/openai-o3-mini/"]},{"id":"qwen2.5-coder-32b-instruct","modelId":"model:alibaba/qwen2.5-coder-32b-instruct","label":"Qwen2.5 Coder 32B Instruct","configuration":"Qwen2.5 Coder 32B Instruct · qwen2.5-coder-32b-instruct","value":46.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct"]},{"id":"gemini-2.0-flash-lite","modelId":"model:google/gemini-2.0-flash-lite","label":"Gemini 2.0 Flash Lite","configuration":"Gemini 2.0 Flash Lite · gemini-2.0-flash-lite","value":58.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://developers.googleblog.com/en/start-building-with-the-gemini-2-0-flash-family/"]},{"id":"qwq-32b","modelId":"model:alibaba/qwq-32b","label":"QwQ 32B","configuration":"QwQ 32B · qwq-32b","value":77.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://qwenlm.github.io/blog/qwq-32b/"]},{"id":"gemma-3-27b-it","modelId":"model:google/gemma-3-27b-it","label":"Gemma 3 27B","configuration":"Gemma 3 27B · gemma-3-27b-it","value":55.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/developers/gemma-3/"]},{"id":"deepseek-v3-0324","modelId":"model:deepseek/deepseek-v3-0324","label":"DeepSeek V3 0324","configuration":"DeepSeek V3 0324 · deepseek-v3-0324","value":73.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3-0324"]}]},{"id":"livebench-data-2024-11-25","name":"LiveBench · 数据分析 · 2024-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:61806df0681ac4c65b21e995447d2fcf35020adf0e14a3f6d9d8d68146f4df4f:17ca6b3d122921d24b591d0da52566c5552ed4f4001ae9d83a11b2712325aae4","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2024-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 19 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"amazon.nova-lite-v1:0","modelId":"model:amazon/nova-lite","label":"Nova Lite","configuration":"Nova Lite · amazon.nova-lite-v1:0","value":37.23,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-micro-v1:0","modelId":"model:amazon/nova-micro","label":"Nova Micro","configuration":"Nova Micro · amazon.nova-micro-v1:0","value":33.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-pro-v1:0","modelId":"model:amazon/nova-pro","label":"Nova Pro","configuration":"Nova Pro · amazon.nova-pro-v1:0","value":48.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"claude-3-5-haiku-20241022","modelId":"model:anthropic/claude-3-5-haiku-20241022","label":"Claude 3.5 Haiku","configuration":"Claude 3.5 Haiku · claude-3-5-haiku-20241022","value":48.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/claude/haiku"]},{"id":"claude-3-5-sonnet-20241022","modelId":"model:anthropic/claude-3-5-sonnet-20241022","label":"Claude 3.5 Sonnet","configuration":"Claude 3.5 Sonnet · claude-3-5-sonnet-20241022","value":55.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/news/3-5-models-and-computer-use"]},{"id":"command-r-plus-08-2024","modelId":"model:cohere/command-r-plus-08-2024","label":"Command R Plus","configuration":"Command R Plus · command-r-plus-08-2024","value":38.06,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://docs.cohere.com/docs/models"]},{"id":"deepseek-r1","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1","value":69.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1"]},{"id":"deepseek-r1-distill-qwen-32b","modelId":"model:deepseek/deepseek-r1-distill-qwen-32b","label":"DeepSeek R1 Distill Qwen 32B","configuration":"DeepSeek R1 Distill Qwen 32B · deepseek-r1-distill-qwen-32b","value":45.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"]},{"id":"deepseek-v3","modelId":"model:deepseek/deepseek-v3","label":"DeepSeek V3","configuration":"DeepSeek V3 · deepseek-v3","value":60.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://api-docs.deepseek.com/news/news1226"]},{"id":"gemini-2.0-flash","modelId":"model:google/gemini-2.0-flash","label":"Gemini 2.0 Flash","configuration":"Gemini 2.0 Flash · gemini-2.0-flash","value":67.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/"]},{"id":"gpt-4o-mini-2024-07-18","modelId":"model:openai/gpt-4o-mini","label":"GPT-4o Mini","configuration":"GPT-4o Mini · gpt-4o-mini-2024-07-18","value":49.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/hello-gpt-4o/"]},{"id":"mistral-large-2411","modelId":"model:mistral/mistral-large-2411","label":"Mistral Large","configuration":"Mistral Large · mistral-large-2411","value":50.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/mistralai/Mistral-Large-Instruct-2411"]},{"id":"o1-2024-12-17-high","modelId":"model:openai/o1","label":"o1 High","configuration":"o1 High · o1-2024-12-17-high","value":65.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/o1/"]},{"id":"o3-mini-2025-01-31-high","modelId":"model:openai/o3-mini","label":"o3-Mini High","configuration":"o3-Mini High · o3-mini-2025-01-31-high","value":70.64,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/openai-o3-mini/"]},{"id":"qwen2.5-coder-32b-instruct","modelId":"model:alibaba/qwen2.5-coder-32b-instruct","label":"Qwen2.5 Coder 32B Instruct","configuration":"Qwen2.5 Coder 32B Instruct · qwen2.5-coder-32b-instruct","value":49.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct"]},{"id":"gemini-2.0-flash-lite","modelId":"model:google/gemini-2.0-flash-lite","label":"Gemini 2.0 Flash Lite","configuration":"Gemini 2.0 Flash Lite · gemini-2.0-flash-lite","value":65.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://developers.googleblog.com/en/start-building-with-the-gemini-2-0-flash-family/"]},{"id":"qwq-32b","modelId":"model:alibaba/qwq-32b","label":"QwQ 32B","configuration":"QwQ 32B · qwq-32b","value":65.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://qwenlm.github.io/blog/qwq-32b/"]},{"id":"gemma-3-27b-it","modelId":"model:google/gemma-3-27b-it","label":"Gemma 3 27B","configuration":"Gemma 3 27B · gemma-3-27b-it","value":51.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/developers/gemma-3/"]},{"id":"deepseek-v3-0324","modelId":"model:deepseek/deepseek-v3-0324","label":"DeepSeek V3 0324","configuration":"DeepSeek V3 0324 · deepseek-v3-0324","value":60.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3-0324"]}]},{"id":"livebench-language-2024-11-25","name":"LiveBench · 语言 · 2024-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:61806df0681ac4c65b21e995447d2fcf35020adf0e14a3f6d9d8d68146f4df4f:17ca6b3d122921d24b591d0da52566c5552ed4f4001ae9d83a11b2712325aae4","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2024-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 19 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"amazon.nova-lite-v1:0","modelId":"model:amazon/nova-lite","label":"Nova Lite","configuration":"Nova Lite · amazon.nova-lite-v1:0","value":25.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-micro-v1:0","modelId":"model:amazon/nova-micro","label":"Nova Micro","configuration":"Nova Micro · amazon.nova-micro-v1:0","value":15.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-pro-v1:0","modelId":"model:amazon/nova-pro","label":"Nova Pro","configuration":"Nova Pro · amazon.nova-pro-v1:0","value":36.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"claude-3-5-haiku-20241022","modelId":"model:anthropic/claude-3-5-haiku-20241022","label":"Claude 3.5 Haiku","configuration":"Claude 3.5 Haiku · claude-3-5-haiku-20241022","value":35.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/claude/haiku"]},{"id":"claude-3-5-sonnet-20241022","modelId":"model:anthropic/claude-3-5-sonnet-20241022","label":"Claude 3.5 Sonnet","configuration":"Claude 3.5 Sonnet · claude-3-5-sonnet-20241022","value":53.76,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/news/3-5-models-and-computer-use"]},{"id":"command-r-plus-08-2024","modelId":"model:cohere/command-r-plus-08-2024","label":"Command R Plus","configuration":"Command R Plus · command-r-plus-08-2024","value":29.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://docs.cohere.com/docs/models"]},{"id":"deepseek-r1","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1","value":48.53,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1"]},{"id":"deepseek-r1-distill-qwen-32b","modelId":"model:deepseek/deepseek-r1-distill-qwen-32b","label":"DeepSeek R1 Distill Qwen 32B","configuration":"DeepSeek R1 Distill Qwen 32B · deepseek-r1-distill-qwen-32b","value":26.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"]},{"id":"deepseek-v3","modelId":"model:deepseek/deepseek-v3","label":"DeepSeek V3","configuration":"DeepSeek V3 · deepseek-v3","value":47.48,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://api-docs.deepseek.com/news/news1226"]},{"id":"gemini-2.0-flash","modelId":"model:google/gemini-2.0-flash","label":"Gemini 2.0 Flash","configuration":"Gemini 2.0 Flash · gemini-2.0-flash","value":40.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/"]},{"id":"gpt-4o-mini-2024-07-18","modelId":"model:openai/gpt-4o-mini","label":"GPT-4o Mini","configuration":"GPT-4o Mini · gpt-4o-mini-2024-07-18","value":28.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/hello-gpt-4o/"]},{"id":"mistral-large-2411","modelId":"model:mistral/mistral-large-2411","label":"Mistral Large","configuration":"Mistral Large · mistral-large-2411","value":39.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/mistralai/Mistral-Large-Instruct-2411"]},{"id":"o1-2024-12-17-high","modelId":"model:openai/o1","label":"o1 High","configuration":"o1 High · o1-2024-12-17-high","value":65.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/o1/"]},{"id":"o3-mini-2025-01-31-high","modelId":"model:openai/o3-mini","label":"o3-Mini High","configuration":"o3-Mini High · o3-mini-2025-01-31-high","value":50.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/openai-o3-mini/"]},{"id":"qwen2.5-coder-32b-instruct","modelId":"model:alibaba/qwen2.5-coder-32b-instruct","label":"Qwen2.5 Coder 32B Instruct","configuration":"Qwen2.5 Coder 32B Instruct · qwen2.5-coder-32b-instruct","value":23.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct"]},{"id":"gemini-2.0-flash-lite","modelId":"model:google/gemini-2.0-flash-lite","label":"Gemini 2.0 Flash Lite","configuration":"Gemini 2.0 Flash Lite · gemini-2.0-flash-lite","value":33.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://developers.googleblog.com/en/start-building-with-the-gemini-2-0-flash-family/"]},{"id":"qwq-32b","modelId":"model:alibaba/qwq-32b","label":"QwQ 32B","configuration":"QwQ 32B · qwq-32b","value":51.35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://qwenlm.github.io/blog/qwq-32b/"]},{"id":"gemma-3-27b-it","modelId":"model:google/gemma-3-27b-it","label":"Gemma 3 27B","configuration":"Gemma 3 27B · gemma-3-27b-it","value":34.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/developers/gemma-3/"]},{"id":"deepseek-v3-0324","modelId":"model:deepseek/deepseek-v3-0324","label":"DeepSeek V3 0324","configuration":"DeepSeek V3 0324 · deepseek-v3-0324","value":49.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3-0324"]}]},{"id":"livebench-if-2024-11-25","name":"LiveBench · 指令遵循 · 2024-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:61806df0681ac4c65b21e995447d2fcf35020adf0e14a3f6d9d8d68146f4df4f:17ca6b3d122921d24b591d0da52566c5552ed4f4001ae9d83a11b2712325aae4","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2024-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 19 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"amazon.nova-lite-v1:0","modelId":"model:amazon/nova-lite","label":"Nova Lite","configuration":"Nova Lite · amazon.nova-lite-v1:0","value":54.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-micro-v1:0","modelId":"model:amazon/nova-micro","label":"Nova Micro","configuration":"Nova Micro · amazon.nova-micro-v1:0","value":48.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"amazon.nova-pro-v1:0","modelId":"model:amazon/nova-pro","label":"Nova Pro","configuration":"Nova Pro · amazon.nova-pro-v1:0","value":67.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://aws.amazon.com/ai/generative-ai/nova/"]},{"id":"claude-3-5-haiku-20241022","modelId":"model:anthropic/claude-3-5-haiku-20241022","label":"Claude 3.5 Haiku","configuration":"Claude 3.5 Haiku · claude-3-5-haiku-20241022","value":61.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/claude/haiku"]},{"id":"claude-3-5-sonnet-20241022","modelId":"model:anthropic/claude-3-5-sonnet-20241022","label":"Claude 3.5 Sonnet","configuration":"Claude 3.5 Sonnet · claude-3-5-sonnet-20241022","value":69.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://www.anthropic.com/news/3-5-models-and-computer-use"]},{"id":"command-r-plus-08-2024","modelId":"model:cohere/command-r-plus-08-2024","label":"Command R Plus","configuration":"Command R Plus · command-r-plus-08-2024","value":57.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://docs.cohere.com/docs/models"]},{"id":"deepseek-r1","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1","value":80.51,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1"]},{"id":"deepseek-r1-distill-qwen-32b","modelId":"model:deepseek/deepseek-r1-distill-qwen-32b","label":"DeepSeek R1 Distill Qwen 32B","configuration":"DeepSeek R1 Distill Qwen 32B · deepseek-r1-distill-qwen-32b","value":55.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B"]},{"id":"deepseek-v3","modelId":"model:deepseek/deepseek-v3","label":"DeepSeek V3","configuration":"DeepSeek V3 · deepseek-v3","value":75.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://api-docs.deepseek.com/news/news1226"]},{"id":"gemini-2.0-flash","modelId":"model:google/gemini-2.0-flash","label":"Gemini 2.0 Flash","configuration":"Gemini 2.0 Flash · gemini-2.0-flash","value":85.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/google-deepmind/gemini-model-updates-february-2025/"]},{"id":"gpt-4o-mini-2024-07-18","modelId":"model:openai/gpt-4o-mini","label":"GPT-4o Mini","configuration":"GPT-4o Mini · gpt-4o-mini-2024-07-18","value":56.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/hello-gpt-4o/"]},{"id":"mistral-large-2411","modelId":"model:mistral/mistral-large-2411","label":"Mistral Large","configuration":"Mistral Large · mistral-large-2411","value":67.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/mistralai/Mistral-Large-Instruct-2411"]},{"id":"o1-2024-12-17-high","modelId":"model:openai/o1","label":"o1 High","configuration":"o1 High · o1-2024-12-17-high","value":81.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/o1/"]},{"id":"o3-mini-2025-01-31-high","modelId":"model:openai/o3-mini","label":"o3-Mini High","configuration":"o3-Mini High · o3-mini-2025-01-31-high","value":84.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://openai.com/index/openai-o3-mini/"]},{"id":"qwen2.5-coder-32b-instruct","modelId":"model:alibaba/qwen2.5-coder-32b-instruct","label":"Qwen2.5 Coder 32B Instruct","configuration":"Qwen2.5 Coder 32B Instruct · qwen2.5-coder-32b-instruct","value":58.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct"]},{"id":"gemini-2.0-flash-lite","modelId":"model:google/gemini-2.0-flash-lite","label":"Gemini 2.0 Flash Lite","configuration":"Gemini 2.0 Flash Lite · gemini-2.0-flash-lite","value":76.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://developers.googleblog.com/en/start-building-with-the-gemini-2-0-flash-family/"]},{"id":"qwq-32b","modelId":"model:alibaba/qwq-32b","label":"QwQ 32B","configuration":"QwQ 32B · qwq-32b","value":81.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://qwenlm.github.io/blog/qwq-32b/"]},{"id":"gemma-3-27b-it","modelId":"model:google/gemma-3-27b-it","label":"Gemma 3 27B","configuration":"Gemma 3 27B · gemma-3-27b-it","value":74.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://blog.google/technology/developers/gemma-3/"]},{"id":"deepseek-v3-0324","modelId":"model:deepseek/deepseek-v3-0324","label":"DeepSeek V3 0324","configuration":"DeepSeek V3 0324 · deepseek-v3-0324","value":81.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2024_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3-0324"]}]},{"id":"livebench-overall-2025-05-30","name":"LiveBench · 全科 · 2025-05-30","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:e733acb6ff38039109f04386a1f8cf91e7f0295daefe4a8e544cde90e03950e5:c1215f3df5b2e0a5fe69907c101d29922bf783b011df9ba82638c0145cb4c64c","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-05-30","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 24 个型号 / 官方材料 70 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":70.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"command-a-03-2025","modelId":"model:cohere/command-a-03-2025","label":"Command A","configuration":"Command A · command-a-03-2025","value":44.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://cohere.com/blog/command-a"]},{"id":"deepseek-r1-0528","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1-0528","value":69.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"]},{"id":"gpt-5-high","modelId":"model:openai/gpt-5","label":"GPT-5 High","configuration":"GPT-5 High · gpt-5-high","value":79.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-codex","modelId":"model:openai/gpt-5-codex","label":"GPT-5 Codex","configuration":"GPT-5 Codex · gpt-5-codex","value":78.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5-codex"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":75.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":78.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":55.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"mistral-medium-2505","modelId":"model:mistral/mistral-medium-2505","label":"Mistral Medium 3","configuration":"Mistral Medium 3 · mistral-medium-2505","value":50.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://mistral.ai/news/mistral-medium-3"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":65.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":57.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":62.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-coder-480b-a35b-instruct","modelId":"model:alibaba/qwen3-coder-480b-a35b-instruct","label":"Qwen 3 Coder 480B A35B Instruct","configuration":"Qwen 3 Coder 480B A35B Instruct · qwen3-coder-480b-a35b-instruct","value":61.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3-coder/"]},{"id":"qwen3-max-2025-09-23","modelId":"model:alibaba/qwen3-max","label":"Qwen 3 Max","configuration":"Qwen 3 Max · qwen3-max-2025-09-23","value":69.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=241398b9cd6353de490b0f82806c7848c5d2777d&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":66.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":64.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":60.42,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":71.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":78.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":75.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":70.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":79.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"claude-opus-4-5-20251101-medium-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Medium Effort","configuration":"Claude 4.5 Opus Medium Effort · claude-opus-4-5-20251101-medium-effort","value":75.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-opus-4-5"]}]},{"id":"livebench-reasoning-2025-05-30","name":"LiveBench · 推理 · 2025-05-30","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:e733acb6ff38039109f04386a1f8cf91e7f0295daefe4a8e544cde90e03950e5:c1215f3df5b2e0a5fe69907c101d29922bf783b011df9ba82638c0145cb4c64c","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-05-30","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 24 个型号 / 官方材料 70 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":63.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"command-a-03-2025","modelId":"model:cohere/command-a-03-2025","label":"Command A","configuration":"Command A · command-a-03-2025","value":36.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://cohere.com/blog/command-a"]},{"id":"deepseek-r1-0528","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1-0528","value":91.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"]},{"id":"gpt-5-high","modelId":"model:openai/gpt-5","label":"GPT-5 High","configuration":"GPT-5 High · gpt-5-high","value":98.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-codex","modelId":"model:openai/gpt-5-codex","label":"GPT-5 Codex","configuration":"GPT-5 Codex · gpt-5-codex","value":98.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5-codex"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":91.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":64.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":96.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":77.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"mistral-medium-2505","modelId":"model:mistral/mistral-medium-2505","label":"Mistral Medium 3","configuration":"Mistral Medium 3 · mistral-medium-2505","value":41.97,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://mistral.ai/news/mistral-medium-3"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":86.89,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":71.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":83.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-coder-480b-a35b-instruct","modelId":"model:alibaba/qwen3-coder-480b-a35b-instruct","label":"Qwen 3 Coder 480B A35B Instruct","configuration":"Qwen 3 Coder 480B A35B Instruct · qwen3-coder-480b-a35b-instruct","value":54.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3-coder/"]},{"id":"qwen3-max-2025-09-23","modelId":"model:alibaba/qwen3-max","label":"Qwen 3 Max","configuration":"Qwen 3 Max · qwen3-max-2025-09-23","value":89,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=241398b9cd6353de490b0f82806c7848c5d2777d&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":88.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":91.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":54.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":87.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":95.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":89.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":98.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"claude-opus-4-5-20251101-medium-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Medium Effort","configuration":"Claude 4.5 Opus Medium Effort · claude-opus-4-5-20251101-medium-effort","value":72.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-opus-4-5"]}]},{"id":"livebench-coding-2025-05-30","name":"LiveBench · 编程答题 · 2025-05-30","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:e733acb6ff38039109f04386a1f8cf91e7f0295daefe4a8e544cde90e03950e5:c1215f3df5b2e0a5fe69907c101d29922bf783b011df9ba82638c0145cb4c64c","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-05-30","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 24 个型号 / 官方材料 70 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":76.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"command-a-03-2025","modelId":"model:cohere/command-a-03-2025","label":"Command A","configuration":"Command A · command-a-03-2025","value":55.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://cohere.com/blog/command-a"]},{"id":"deepseek-r1-0528","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1-0528","value":73.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"]},{"id":"gpt-5-high","modelId":"model:openai/gpt-5","label":"GPT-5 High","configuration":"GPT-5 High · gpt-5-high","value":77.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-codex","modelId":"model:openai/gpt-5-codex","label":"GPT-5 Codex","configuration":"GPT-5 Codex · gpt-5-codex","value":69.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5-codex"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":67.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":72.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":60.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"mistral-medium-2505","modelId":"model:mistral/mistral-medium-2505","label":"Mistral Medium 3","configuration":"Mistral Medium 3 · mistral-medium-2505","value":63.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://mistral.ai/news/mistral-medium-3"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":69.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":48.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":66.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-coder-480b-a35b-instruct","modelId":"model:alibaba/qwen3-coder-480b-a35b-instruct","label":"Qwen 3 Coder 480B A35B Instruct","configuration":"Qwen 3 Coder 480B A35B Instruct · qwen3-coder-480b-a35b-instruct","value":74.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3-coder/"]},{"id":"qwen3-max-2025-09-23","modelId":"model:alibaba/qwen3-max","label":"Qwen 3 Max","configuration":"Qwen 3 Max · qwen3-max-2025-09-23","value":71.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=241398b9cd6353de490b0f82806c7848c5d2777d&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":60.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":72.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":72.49,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":71.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":69.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":74.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"claude-opus-4-5-20251101-medium-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Medium Effort","configuration":"Claude 4.5 Opus Medium Effort · claude-opus-4-5-20251101-medium-effort","value":78.51,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-opus-4-5"]}]},{"id":"livebench-agentic-2025-05-30","name":"LiveBench · 代理编程 · 2025-05-30","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:e733acb6ff38039109f04386a1f8cf91e7f0295daefe4a8e544cde90e03950e5:c1215f3df5b2e0a5fe69907c101d29922bf783b011df9ba82638c0145cb4c64c","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-05-30","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 24 个型号 / 官方材料 70 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":50,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"command-a-03-2025","modelId":"model:cohere/command-a-03-2025","label":"Command A","configuration":"Command A · command-a-03-2025","value":3.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://cohere.com/blog/command-a"]},{"id":"deepseek-r1-0528","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1-0528","value":20,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"]},{"id":"gpt-5-high","modelId":"model:openai/gpt-5","label":"GPT-5 High","configuration":"GPT-5 High · gpt-5-high","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-codex","modelId":"model:openai/gpt-5-codex","label":"GPT-5 Codex","configuration":"GPT-5 Codex · gpt-5-codex","value":48.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5-codex"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":43.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":23.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":43.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"mistral-medium-2505","modelId":"model:mistral/mistral-medium-2505","label":"Mistral Medium 3","configuration":"Mistral Medium 3 · mistral-medium-2505","value":10,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://mistral.ai/news/mistral-medium-3"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":13.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":1.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":1.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-coder-480b-a35b-instruct","modelId":"model:alibaba/qwen3-coder-480b-a35b-instruct","label":"Qwen 3 Coder 480B A35B Instruct","configuration":"Qwen 3 Coder 480B A35B Instruct · qwen3-coder-480b-a35b-instruct","value":31.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3-coder/"]},{"id":"qwen3-max-2025-09-23","modelId":"model:alibaba/qwen3-max","label":"Qwen 3 Max","configuration":"Qwen 3 Max · qwen3-max-2025-09-23","value":31.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=241398b9cd6353de490b0f82806c7848c5d2777d&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":10,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":28.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":43.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":43.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"claude-opus-4-5-20251101-medium-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Medium Effort","configuration":"Claude 4.5 Opus Medium Effort · claude-opus-4-5-20251101-medium-effort","value":63.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-opus-4-5"]}]},{"id":"livebench-math-2025-05-30","name":"LiveBench · 数学 · 2025-05-30","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:e733acb6ff38039109f04386a1f8cf91e7f0295daefe4a8e544cde90e03950e5:c1215f3df5b2e0a5fe69907c101d29922bf783b011df9ba82638c0145cb4c64c","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-05-30","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 24 个型号 / 官方材料 70 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":82.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"command-a-03-2025","modelId":"model:cohere/command-a-03-2025","label":"Command A","configuration":"Command A · command-a-03-2025","value":45.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://cohere.com/blog/command-a"]},{"id":"deepseek-r1-0528","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1-0528","value":85.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"]},{"id":"gpt-5-high","modelId":"model:openai/gpt-5","label":"GPT-5 High","configuration":"GPT-5 High · gpt-5-high","value":92.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-codex","modelId":"model:openai/gpt-5-codex","label":"GPT-5 Codex","configuration":"GPT-5 Codex · gpt-5-codex","value":92.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5-codex"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":90.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":71.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":93.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":69.89,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"mistral-medium-2505","modelId":"model:mistral/mistral-medium-2505","label":"Mistral Medium 3","configuration":"Mistral Medium 3 · mistral-medium-2505","value":59.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://mistral.ai/news/mistral-medium-3"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":79.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":76.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":80.05,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-coder-480b-a35b-instruct","modelId":"model:alibaba/qwen3-coder-480b-a35b-instruct","label":"Qwen 3 Coder 480B A35B Instruct","configuration":"Qwen 3 Coder 480B A35B Instruct · qwen3-coder-480b-a35b-instruct","value":67.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3-coder/"]},{"id":"qwen3-max-2025-09-23","modelId":"model:alibaba/qwen3-max","label":"Qwen 3 Max","configuration":"Qwen 3 Max · qwen3-max-2025-09-23","value":83.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=241398b9cd6353de490b0f82806c7848c5d2777d&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":80.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":82.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":74.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":88.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":94.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":87.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":86.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":94.97,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"claude-opus-4-5-20251101-medium-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Medium Effort","configuration":"Claude 4.5 Opus Medium Effort · claude-opus-4-5-20251101-medium-effort","value":87.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-opus-4-5"]}]},{"id":"livebench-data-2025-05-30","name":"LiveBench · 数据分析 · 2025-05-30","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:e733acb6ff38039109f04386a1f8cf91e7f0295daefe4a8e544cde90e03950e5:c1215f3df5b2e0a5fe69907c101d29922bf783b011df9ba82638c0145cb4c64c","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-05-30","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 24 个型号 / 官方材料 70 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":67.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"command-a-03-2025","modelId":"model:cohere/command-a-03-2025","label":"Command A","configuration":"Command A · command-a-03-2025","value":48.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://cohere.com/blog/command-a"]},{"id":"deepseek-r1-0528","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1-0528","value":71.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"]},{"id":"gpt-5-high","modelId":"model:openai/gpt-5","label":"GPT-5 High","configuration":"GPT-5 High · gpt-5-high","value":71.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-codex","modelId":"model:openai/gpt-5-codex","label":"GPT-5 Codex","configuration":"GPT-5 Codex · gpt-5-codex","value":70.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5-codex"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":71.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":65.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":72.42,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":56.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"mistral-medium-2505","modelId":"model:mistral/mistral-medium-2505","label":"Mistral Medium 3","configuration":"Mistral Medium 3 · mistral-medium-2505","value":60.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://mistral.ai/news/mistral-medium-3"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":65.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":66.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":68.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-coder-480b-a35b-instruct","modelId":"model:alibaba/qwen3-coder-480b-a35b-instruct","label":"Qwen 3 Coder 480B A35B Instruct","configuration":"Qwen 3 Coder 480B A35B Instruct · qwen3-coder-480b-a35b-instruct","value":64.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3-coder/"]},{"id":"qwen3-max-2025-09-23","modelId":"model:alibaba/qwen3-max","label":"Qwen 3 Max","configuration":"Qwen 3 Max · qwen3-max-2025-09-23","value":65.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=241398b9cd6353de490b0f82806c7848c5d2777d&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":68.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":73.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":66.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":69.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":72.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":68.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":70.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":74.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"claude-opus-4-5-20251101-medium-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Medium Effort","configuration":"Claude 4.5 Opus Medium Effort · claude-opus-4-5-20251101-medium-effort","value":67.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-opus-4-5"]}]},{"id":"livebench-language-2025-05-30","name":"LiveBench · 语言 · 2025-05-30","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:e733acb6ff38039109f04386a1f8cf91e7f0295daefe4a8e544cde90e03950e5:c1215f3df5b2e0a5fe69907c101d29922bf783b011df9ba82638c0145cb4c64c","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-05-30","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 24 个型号 / 官方材料 70 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":76.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"command-a-03-2025","modelId":"model:cohere/command-a-03-2025","label":"Command A","configuration":"Command A · command-a-03-2025","value":36.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://cohere.com/blog/command-a"]},{"id":"deepseek-r1-0528","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1-0528","value":64.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"]},{"id":"gpt-5-high","modelId":"model:openai/gpt-5","label":"GPT-5 High","configuration":"GPT-5 High · gpt-5-high","value":80.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-codex","modelId":"model:openai/gpt-5-codex","label":"GPT-5 Codex","configuration":"GPT-5 Codex · gpt-5-codex","value":79.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5-codex"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":75.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":46.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":81.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":43.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"mistral-medium-2505","modelId":"model:mistral/mistral-medium-2505","label":"Mistral Medium 3","configuration":"Mistral Medium 3 · mistral-medium-2505","value":44.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://mistral.ai/news/mistral-medium-3"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":66.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":56.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":55.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-coder-480b-a35b-instruct","modelId":"model:alibaba/qwen3-coder-480b-a35b-instruct","label":"Qwen 3 Coder 480B A35B Instruct","configuration":"Qwen 3 Coder 480B A35B Instruct · qwen3-coder-480b-a35b-instruct","value":64.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3-coder/"]},{"id":"qwen3-max-2025-09-23","modelId":"model:alibaba/qwen3-max","label":"Qwen 3 Max","configuration":"Qwen 3 Max · qwen3-max-2025-09-23","value":71.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=241398b9cd6353de490b0f82806c7848c5d2777d&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":68.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":54.48,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":53.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":61.92,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":80.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":70.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":61.4,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":83.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"claude-opus-4-5-20251101-medium-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Medium Effort","configuration":"Claude 4.5 Opus Medium Effort · claude-opus-4-5-20251101-medium-effort","value":79.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-opus-4-5"]}]},{"id":"livebench-if-2025-05-30","name":"LiveBench · 指令遵循 · 2025-05-30","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:e733acb6ff38039109f04386a1f8cf91e7f0295daefe4a8e544cde90e03950e5:c1215f3df5b2e0a5fe69907c101d29922bf783b011df9ba82638c0145cb4c64c","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-05-30","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 24 个型号 / 官方材料 70 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":79.05,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"command-a-03-2025","modelId":"model:cohere/command-a-03-2025","label":"Command A","configuration":"Command A · command-a-03-2025","value":82.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://cohere.com/blog/command-a"]},{"id":"deepseek-r1-0528","modelId":"model:deepseek/deepseek-r1","label":"DeepSeek R1","configuration":"DeepSeek R1 · deepseek-r1-0528","value":79.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/deepseek-ai/DeepSeek-R1-0528"]},{"id":"gpt-5-high","modelId":"model:openai/gpt-5","label":"GPT-5 High","configuration":"GPT-5 High · gpt-5-high","value":88.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-codex","modelId":"model:openai/gpt-5-codex","label":"GPT-5 Codex","configuration":"GPT-5 Codex · gpt-5-codex","value":88.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5-codex"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":85.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":74.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":91.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":66.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"mistral-medium-2505","modelId":"model:mistral/mistral-medium-2505","label":"Mistral Medium 3","configuration":"Mistral Medium 3 · mistral-medium-2505","value":71.4,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://mistral.ai/news/mistral-medium-3"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":75.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":83.23,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":85.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-coder-480b-a35b-instruct","modelId":"model:alibaba/qwen3-coder-480b-a35b-instruct","label":"Qwen 3 Coder 480B A35B Instruct","configuration":"Qwen 3 Coder 480B A35B Instruct · qwen3-coder-480b-a35b-instruct","value":74.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwenlm.github.io/blog/qwen3-coder/"]},{"id":"qwen3-max-2025-09-23","modelId":"model:alibaba/qwen3-max","label":"Qwen 3 Max","configuration":"Qwen 3 Max · qwen3-max-2025-09-23","value":76.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=241398b9cd6353de490b0f82806c7848c5d2777d&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":80.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":85.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":73.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":90.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":93.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":85.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":80.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":85.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"claude-opus-4-5-20251101-medium-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Medium Effort","configuration":"Claude 4.5 Opus Medium Effort · claude-opus-4-5-20251101-medium-effort","value":81.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_05_30.csv","https://www.anthropic.com/news/claude-opus-4-5"]}]},{"id":"livebench-overall-2025-11-25","name":"LiveBench · 全科 · 2025-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:a3f8569ce5b6e0a843d8d160cce2dcba209e5c035bb38feac6dfd89e80037d12:34242000938912ac33369e3b25d49f6eafd3cc64823914b2405cef9a7ddcd81d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":76.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":59.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":57.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":46.51,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":73.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":75.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":69.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":54.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":74.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":73.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":72.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":76.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":65.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":74.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":51.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":65.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":53.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":45.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":49.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":53.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":54.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-reasoning-2025-11-25","name":"LiveBench · 推理 · 2025-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:a3f8569ce5b6e0a843d8d160cce2dcba209e5c035bb38feac6dfd89e80037d12:34242000938912ac33369e3b25d49f6eafd3cc64823914b2405cef9a7ddcd81d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":39.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":84.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":45.72,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":31.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":75.4,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":79.89,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":70.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":44.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":87.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":83.06,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":87.97,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":91.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":66.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":86.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":49.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":65.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":57.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":45.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":54.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":55.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":58.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-coding-2025-11-25","name":"LiveBench · 编程答题 · 2025-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:a3f8569ce5b6e0a843d8d160cce2dcba209e5c035bb38feac6dfd89e80037d12:34242000938912ac33369e3b25d49f6eafd3cc64823914b2405cef9a7ddcd81d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":72.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":79.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":76.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":75.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":66.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":73.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":74.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":67.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":72.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":72.49,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":71.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":81.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":69.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":76.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":60.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":67.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":69.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":48.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":66.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":60.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-agentic-2025-11-25","name":"LiveBench · 代理编程 · 2025-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:a3f8569ce5b6e0a843d8d160cce2dcba209e5c035bb38feac6dfd89e80037d12:34242000938912ac33369e3b25d49f6eafd3cc64823914b2405cef9a7ddcd81d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":33.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":63.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":48.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":43.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":40,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":28.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":53.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":53.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":56.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":40,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":16.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":38.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":13.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":1.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":3.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":10,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":8.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-math-2025-11-25","name":"LiveBench · 数学 · 2025-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:a3f8569ce5b6e0a843d8d160cce2dcba209e5c035bb38feac6dfd89e80037d12:34242000938912ac33369e3b25d49f6eafd3cc64823914b2405cef9a7ddcd81d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":75.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":94.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":80.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":81.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":67.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":93.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":94.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":90.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":78.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":94.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":93.53,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":91.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":92.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":91.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":93.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":81.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":90.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":83.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":79.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":81.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":84.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":86.35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-data-2025-11-25","name":"LiveBench · 数据分析 · 2025-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:a3f8569ce5b6e0a843d8d160cce2dcba209e5c035bb38feac6dfd89e80037d12:34242000938912ac33369e3b25d49f6eafd3cc64823914b2405cef9a7ddcd81d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":66.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":71.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":67.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":66.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":56.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":74.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":74.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":71.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":65.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":72.42,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":72.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":68.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":71.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":70.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":72.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":56.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":70.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":65.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":66.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":68.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":68.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":73.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-language-2025-11-25","name":"LiveBench · 语言 · 2025-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:a3f8569ce5b6e0a843d8d160cce2dcba209e5c035bb38feac6dfd89e80037d12:34242000938912ac33369e3b25d49f6eafd3cc64823914b2405cef9a7ddcd81d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":57.05,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":81.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":76,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":64.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":45.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":84.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":84.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":75.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":47.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":80.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":79.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":69.48,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":75.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":63.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":79.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":48.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":66.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":66.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":54.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":55.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":66.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":56.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-if-2025-11-25","name":"LiveBench · 指令遵循 · 2025-11-25","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:a3f8569ce5b6e0a843d8d160cce2dcba209e5c035bb38feac6dfd89e80037d12:34242000938912ac33369e3b25d49f6eafd3cc64823914b2405cef9a7ddcd81d","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-11-25","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":17.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":62.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":23.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":23.06,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":13.5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":74.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":65.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":65.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":51.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":63.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":63.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":63.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":67.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":59.02,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":61.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":50.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":62.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":21.72,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":21.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":17.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":19.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":41.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_11_25.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-overall-2025-12-23","name":"LiveBench · 全科 · 2025-12-23","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:6d749178f9beb6dd9ba635333b18cc4814cefa8d88f19b07c648cdde898ebc52:d8a96fd89213f11f144c92846567f8652bc2740511e8b4bc4211f659ff57f670","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-12-23","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":50.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":76.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":59.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":57.51,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":45.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":73.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":75.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":69.51,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":53.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":73.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":73.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":71.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":75.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":65.42,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":74.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":50.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":65.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":54.02,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":44.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":48.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":53.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":54.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-reasoning-2025-12-23","name":"LiveBench · 推理 · 2025-12-23","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:6d749178f9beb6dd9ba635333b18cc4814cefa8d88f19b07c648cdde898ebc52:d8a96fd89213f11f144c92846567f8652bc2740511e8b4bc4211f659ff57f670","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-12-23","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":33.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":80.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":42.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":44.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":27.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":74.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":77.42,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":68.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":35.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":81.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":78.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":81.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":84.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":64.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":83.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":39.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":63.49,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":58.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":36.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":48.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":54.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":58.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-coding-2025-12-23","name":"LiveBench · 编程答题 · 2025-12-23","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:6d749178f9beb6dd9ba635333b18cc4814cefa8d88f19b07c648cdde898ebc52:d8a96fd89213f11f144c92846567f8652bc2740511e8b4bc4211f659ff57f670","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-12-23","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":72.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":79.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":76.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":75.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":66.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":73.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":74.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":67.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":72.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":72.49,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":71.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":81.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":69.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":76.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":60.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":67.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":69.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":48.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":66.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":60.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-agentic-2025-12-23","name":"LiveBench · 代理编程 · 2025-12-23","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:6d749178f9beb6dd9ba635333b18cc4814cefa8d88f19b07c648cdde898ebc52:d8a96fd89213f11f144c92846567f8652bc2740511e8b4bc4211f659ff57f670","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-12-23","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":33.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":63.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":48.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":43.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":40,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":28.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":53.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":53.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":56.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":40,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":16.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":38.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":13.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":1.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":3.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":10,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":8.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-math-2025-12-23","name":"LiveBench · 数学 · 2025-12-23","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:6d749178f9beb6dd9ba635333b18cc4814cefa8d88f19b07c648cdde898ebc52:d8a96fd89213f11f144c92846567f8652bc2740511e8b4bc4211f659ff57f670","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-12-23","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":75.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":94.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":80.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":81.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":67.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":93.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":94.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":90.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":78.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":94.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":93.53,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":91.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":92.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":91.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":93.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":81.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":90.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":83.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":79.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":81.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":84.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":86.35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-data-2025-12-23","name":"LiveBench · 数据分析 · 2025-12-23","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:6d749178f9beb6dd9ba635333b18cc4814cefa8d88f19b07c648cdde898ebc52:d8a96fd89213f11f144c92846567f8652bc2740511e8b4bc4211f659ff57f670","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-12-23","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":66.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":71.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":67.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":66.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":56.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":74.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":74.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":71.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":65.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":72.42,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":72.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":68.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":71.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":70.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":72.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":56.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":70.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":65.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":66.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":68.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":68.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":73.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-language-2025-12-23","name":"LiveBench · 语言 · 2025-12-23","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:6d749178f9beb6dd9ba635333b18cc4814cefa8d88f19b07c648cdde898ebc52:d8a96fd89213f11f144c92846567f8652bc2740511e8b4bc4211f659ff57f670","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-12-23","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":57.05,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":81.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":76,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":64.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":45.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":84.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":84.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":75.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":47.73,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":80.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":79.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":69.48,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":75.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":63.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":79.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":48.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":66.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":66.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":54.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":55.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":66.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":56.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-if-2025-12-23","name":"LiveBench · 指令遵循 · 2025-12-23","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:6d749178f9beb6dd9ba635333b18cc4814cefa8d88f19b07c648cdde898ebc52:d8a96fd89213f11f144c92846567f8652bc2740511e8b4bc4211f659ff57f670","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2025-12-23","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 22 个型号 / 官方材料 65 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":17.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":62.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":23.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":23.06,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":13.5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":74.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":65.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":65.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano","configuration":"GPT-5 Nano · gpt-5-nano","value":51.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":63.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":63.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":63.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max","configuration":"GPT-5.1 Codex Max · gpt-5.1-codex-max","value":67.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":59.02,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":61.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":50.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":62.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":21.72,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":21.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":17.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":19.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":41.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2025_12_23.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]}]},{"id":"livebench-overall-2026-01-08","name":"LiveBench · 全科 · 2026-01-08","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:271cb3a3128c5c7ea185988f2c78154281097c87e328824d7527972027f29aae:bcb52d78b463adeccdd06899452a0bf68ffa84906bd37f68fdc0659d9f241241","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-01-08","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 42 个型号 / 官方材料 121 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":45.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":75.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":76.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":53.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":75.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":76.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":51.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":41.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":72.4,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":73.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":79.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":75.02,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gemma-4-31b-it","modelId":"model:google/gemma-4-31b-it","label":"Gemma 4 31B","configuration":"Gemma 4 31B · gemma-4-31b-it","value":61.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":65.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano-high","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano High","configuration":"GPT-5 Nano High · gpt-5-nano-high","value":48.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":70.48,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":72.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":68.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max-high","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max High","configuration":"GPT-5.1 Codex Max High · gpt-5.1-codex-max-high","value":73.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":60.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":74.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":74.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.3-codex-high","modelId":"model:openai/gpt-5.3-codex","label":"GPT-5.3 Codex High","configuration":"GPT-5.3 Codex High · gpt-5.3-codex-high","value":72.76,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.3-codex"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":80.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":67.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":70.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":80.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":46.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":66.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":61.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":72.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"mimo-v2-pro","modelId":"model:xiaomi/mimo-v2-pro","label":"MiMo V2 Pro","configuration":"MiMo V2 Pro · mimo-v2-pro","value":58.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mimo.xiaomi.com/mimo-v2-pro"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":48.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":39.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":43.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":48.35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":50.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":70.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":65.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-flash","modelId":"model:alibaba/qwen3.6-flash","label":"Qwen 3.6 Flash","configuration":"Qwen 3.6 Flash · qwen3.6-flash","value":60.37,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openrouter.ai/qwen/qwen3.6-flash/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":74.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":51.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]}]},{"id":"livebench-reasoning-2026-01-08","name":"LiveBench · 推理 · 2026-01-08","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:271cb3a3128c5c7ea185988f2c78154281097c87e328824d7527972027f29aae:bcb52d78b463adeccdd06899452a0bf68ffa84906bd37f68fdc0659d9f241241","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-01-08","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 42 个型号 / 官方材料 121 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":33.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":80.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":88.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":42.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":84.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":87.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":44.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":27.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":74.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":77.42,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gemma-4-31b-it","modelId":"model:google/gemma-4-31b-it","label":"Gemma 4 31B","configuration":"Gemma 4 31B · gemma-4-31b-it","value":59.42,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":68.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano-high","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano High","configuration":"GPT-5 Nano High · gpt-5-nano-high","value":40.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":81.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":78.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":81.98,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max-high","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max High","configuration":"GPT-5.1 Codex Max High · gpt-5.1-codex-max-high","value":83.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":64.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":83.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":77.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.3-codex-high","modelId":"model:openai/gpt-5.3-codex","label":"GPT-5.3 Codex High","configuration":"GPT-5.3 Codex High · gpt-5.3-codex-high","value":80.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.3-codex"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":88.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":72.5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":81.05,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":87.71,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":39.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":70.82,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":63.49,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":79.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"mimo-v2-pro","modelId":"model:xiaomi/mimo-v2-pro","label":"MiMo V2 Pro","configuration":"MiMo V2 Pro · mimo-v2-pro","value":69.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mimo.xiaomi.com/mimo-v2-pro"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":58.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":36.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":48.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":54.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":58.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":75.83,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":70.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-flash","modelId":"model:alibaba/qwen3.6-flash","label":"Qwen 3.6 Flash","configuration":"Qwen 3.6 Flash · qwen3.6-flash","value":62.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openrouter.ai/qwen/qwen3.6-flash/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":83.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":37.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]}]},{"id":"livebench-coding-2026-01-08","name":"LiveBench · 编程答题 · 2026-01-08","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:271cb3a3128c5c7ea185988f2c78154281097c87e328824d7527972027f29aae:bcb52d78b463adeccdd06899452a0bf68ffa84906bd37f68fdc0659d9f241241","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-01-08","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 42 个型号 / 官方材料 121 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":72.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":79.65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":78.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":76.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":79.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":82.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":75.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":66.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":73.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":74.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":76.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":78.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gemma-4-31b-it","modelId":"model:google/gemma-4-31b-it","label":"Gemma 4 31B","configuration":"Gemma 4 31B · gemma-4-31b-it","value":60.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano-high","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano High","configuration":"GPT-5 Nano High · gpt-5-nano-high","value":62.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":72.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":72.49,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":71.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max-high","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max High","configuration":"GPT-5.1 Codex Max High · gpt-5.1-codex-max-high","value":80.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":69.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":76.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":83.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.3-codex-high","modelId":"model:openai/gpt-5.3-codex","label":"GPT-5.3 Codex High","configuration":"GPT-5.3 Codex High · gpt-5.3-codex-high","value":78.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.3-codex"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":77.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":71.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":72.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":82.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":60.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":69.93,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":67.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":78.57,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"mimo-v2-pro","modelId":"model:xiaomi/mimo-v2-pro","label":"MiMo V2 Pro","configuration":"MiMo V2 Pro · mimo-v2-pro","value":68.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mimo.xiaomi.com/mimo-v2-pro"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":69.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":48.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":66.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":68.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":60.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":78.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":71.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-flash","modelId":"model:alibaba/qwen3.6-flash","label":"Qwen 3.6 Flash","configuration":"Qwen 3.6 Flash · qwen3.6-flash","value":64.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openrouter.ai/qwen/qwen3.6-flash/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":74.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":71.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]}]},{"id":"livebench-agentic-2026-01-08","name":"LiveBench · 代理编程 · 2026-01-08","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:271cb3a3128c5c7ea185988f2c78154281097c87e328824d7527972027f29aae:bcb52d78b463adeccdd06899452a0bf68ffa84906bd37f68fdc0659d9f241241","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-01-08","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 42 个型号 / 官方材料 121 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":33.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":63.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":61.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":48.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":60,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":60,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":43.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":40,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":65,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gemma-4-31b-it","modelId":"model:google/gemma-4-31b-it","label":"Gemma 4 31B","configuration":"Gemma 4 31B · gemma-4-31b-it","value":40,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano-high","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano High","configuration":"GPT-5 Nano High · gpt-5-nano-high","value":23.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":53.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":53.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max-high","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max High","configuration":"GPT-5.1 Codex Max High · gpt-5.1-codex-max-high","value":53.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":40,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.3-codex-high","modelId":"model:openai/gpt-5.3-codex","label":"GPT-5.3 Codex High","configuration":"GPT-5.3 Codex High · gpt-5.3-codex-high","value":55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.3-codex"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":70,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":47.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":49.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":56.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":16.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":50,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":38.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":58.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"mimo-v2-pro","modelId":"model:xiaomi/mimo-v2-pro","label":"MiMo V2 Pro","configuration":"MiMo V2 Pro · mimo-v2-pro","value":30,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mimo.xiaomi.com/mimo-v2-pro"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":13.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":1.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":3.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":10,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":8.33,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":50,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-flash","modelId":"model:alibaba/qwen3.6-flash","label":"Qwen 3.6 Flash","configuration":"Qwen 3.6 Flash · qwen3.6-flash","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openrouter.ai/qwen/qwen3.6-flash/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":51.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":46.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]}]},{"id":"livebench-math-2026-01-08","name":"LiveBench · 数学 · 2026-01-08","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:271cb3a3128c5c7ea185988f2c78154281097c87e328824d7527972027f29aae:bcb52d78b463adeccdd06899452a0bf68ffa84906bd37f68fdc0659d9f241241","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-01-08","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 42 个型号 / 官方材料 121 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":57.97,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":90.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":89.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":62.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":86.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":93.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":63.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":52.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":84.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":81.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":91.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":88.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gemma-4-31b-it","modelId":"model:google/gemma-4-31b-it","label":"Gemma 4 31B","configuration":"Gemma 4 31B · gemma-4-31b-it","value":73.94,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":82.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano-high","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano High","configuration":"GPT-5 Nano High · gpt-5-nano-high","value":68.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":86.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":86.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":79.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max-high","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max High","configuration":"GPT-5.1 Codex Max High · gpt-5.1-codex-max-high","value":83.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":76.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":93.17,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":88.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.3-codex-high","modelId":"model:openai/gpt-5.3-codex","label":"GPT-5.3 Codex High","configuration":"GPT-5.3 Codex High · gpt-5.3-codex-high","value":87.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.3-codex"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":94.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":78.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":91.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":96.32,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":68.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":84.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":81.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":84.28,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"mimo-v2-pro","modelId":"model:xiaomi/mimo-v2-pro","label":"MiMo V2 Pro","configuration":"MiMo V2 Pro · mimo-v2-pro","value":76.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mimo.xiaomi.com/mimo-v2-pro"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":68.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":65.35,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":67.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":70.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":74.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":83.72,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":79.87,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-flash","modelId":"model:alibaba/qwen3.6-flash","label":"Qwen 3.6 Flash","configuration":"Qwen 3.6 Flash · qwen3.6-flash","value":78.88,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openrouter.ai/qwen/qwen3.6-flash/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":85.25,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":54.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]}]},{"id":"livebench-data-2026-01-08","name":"LiveBench · 数据分析 · 2026-01-08","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:271cb3a3128c5c7ea185988f2c78154281097c87e328824d7527972027f29aae:bcb52d78b463adeccdd06899452a0bf68ffa84906bd37f68fdc0659d9f241241","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-01-08","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 42 个型号 / 官方材料 121 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":45.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":74.44,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":69.89,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":77.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":78.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":45.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":39.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":74.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":74.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":78.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":64.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gemma-4-31b-it","modelId":"model:google/gemma-4-31b-it","label":"Gemma 4 31B","configuration":"Gemma 4 31B · gemma-4-31b-it","value":58.76,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":55.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano-high","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano High","configuration":"GPT-5 Nano High · gpt-5-nano-high","value":43.41,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":57.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":69.61,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":60.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max-high","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max High","configuration":"GPT-5.1 Codex Max High · gpt-5.1-codex-max-high","value":70.12,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":49.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":78.16,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":78.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.3-codex-high","modelId":"model:openai/gpt-5.3-codex","label":"GPT-5.3 Codex High","configuration":"GPT-5.3 Codex High · gpt-5.3-codex-high","value":62.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.3-codex"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":79.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":70.95,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":67.64,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":81.08,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":38.8,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":55.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":52.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":65.13,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"mimo-v2-pro","modelId":"model:xiaomi/mimo-v2-pro","label":"MiMo V2 Pro","configuration":"MiMo V2 Pro · mimo-v2-pro","value":49.21,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mimo.xiaomi.com/mimo-v2-pro"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":44.72,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":44.92,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":46.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":49.78,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":53.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":69.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":70.43,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-flash","modelId":"model:alibaba/qwen3.6-flash","label":"Qwen 3.6 Flash","configuration":"Qwen 3.6 Flash · qwen3.6-flash","value":58.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openrouter.ai/qwen/qwen3.6-flash/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":71.79,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":42.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]}]},{"id":"livebench-language-2026-01-08","name":"LiveBench · 语言 · 2026-01-08","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:271cb3a3128c5c7ea185988f2c78154281097c87e328824d7527972027f29aae:bcb52d78b463adeccdd06899452a0bf68ffa84906bd37f68fdc0659d9f241241","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-01-08","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 42 个型号 / 官方材料 121 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":57.05,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":81.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":83.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":76,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":76.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":77.91,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":64.24,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":45.67,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":84.56,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":84.62,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":85.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":84.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gemma-4-31b-it","modelId":"model:google/gemma-4-31b-it","label":"Gemma 4 31B","configuration":"Gemma 4 31B · gemma-4-31b-it","value":71.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":75.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano-high","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano High","configuration":"GPT-5 Nano High · gpt-5-nano-high","value":46.84,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":80.69,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":79.26,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":69.48,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max-high","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max High","configuration":"GPT-5.1 Codex Max High · gpt-5.1-codex-max-high","value":76.48,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":63.01,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":79.81,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":73.68,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.3-codex-high","modelId":"model:openai/gpt-5.3-codex","label":"GPT-5.3 Codex High","configuration":"GPT-5.3 Codex High · gpt-5.3-codex-high","value":80.09,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.3-codex"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":82.63,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":71.46,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":62.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":87.66,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":48.59,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":73.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":66.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":75.14,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"mimo-v2-pro","modelId":"model:xiaomi/mimo-v2-pro","label":"MiMo V2 Pro","configuration":"MiMo V2 Pro · mimo-v2-pro","value":69.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mimo.xiaomi.com/mimo-v2-pro"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":66.07,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":54.47,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":55.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":66.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":56.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":74.99,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":63.3,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-flash","modelId":"model:alibaba/qwen3.6-flash","label":"Qwen 3.6 Flash","configuration":"Qwen 3.6 Flash · qwen3.6-flash","value":63.15,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openrouter.ai/qwen/qwen3.6-flash/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":79.74,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":52.18,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]}]},{"id":"livebench-if-2026-01-08","name":"LiveBench · 指令遵循 · 2026-01-08","kind":"benchmark","sourceId":"livebench","sourceUrl":"https://livebench.ai/","sourceVersion":"livebench:271cb3a3128c5c7ea185988f2c78154281097c87e328824d7527972027f29aae:bcb52d78b463adeccdd06899452a0bf68ffa84906bd37f68fdc0659d9f241241","sourcePublishedAt":"2026-09-29T19:01:46.000Z","verifiedAt":"2026-09-30T02:15:05.886Z","expiresAt":"2026-10-29T00:00:00.000Z","methodology":"来源发布榜：类别内任务均分，全科为类别均分；显示保留两位小数。各行配置不同，不是本站控制变量复测。实测日期、运行脚手架修订和服务配置未公开核验，不生成 BenchmarkRun 或能力—成本前沿。","dataset":"LiveBench 2026-01-08","selectionPolicy":"显式型号映射；排除微调衍生项；每型号保留全科最高配置（两位小数），同分按来源 ID 字典序；所有子榜沿用该配置，不另挑单科最高。缺任务不补零，缺任一类别不进入全科选择。","scope":"本站明确映射的 42 个型号 / 官方材料 121 个配置行；这是收录子集排名，并非官网完整名次。","ruleVersion":"arena-v2.1","fitVersion":null,"unit":"points","higherIsBetter":true,"scale":{"kind":"bounded","min":0,"max":100},"costBoundary":null,"rows":[{"id":"claude-haiku-4-5-20251001","modelId":"model:anthropic/claude-haiku-4-5-20251001","label":"Claude Haiku 4.5","configuration":"Claude Haiku 4.5 · claude-haiku-4-5-20251001","value":17.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-haiku-4-5"]},{"id":"claude-opus-4-5-20251101-thinking-64k-high-effort","modelId":"model:anthropic/claude-opus-4-5-20251101","label":"Claude 4.5 Opus Thinking High Effort","configuration":"Claude 4.5 Opus Thinking High Effort · claude-opus-4-5-20251101-thinking-64k-high-effort","value":62.55,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-5"]},{"id":"claude-opus-4-6-thinking-auto-high-effort","modelId":"model:anthropic/claude-opus-4-6","label":"Claude 4.6 Opus Thinking High Effort","configuration":"Claude 4.6 Opus Thinking High Effort · claude-opus-4-6-thinking-auto-high-effort","value":63.31,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-6"]},{"id":"claude-sonnet-4-5-20250929","modelId":"model:anthropic/claude-sonnet-4-5-20250929","label":"Claude Sonnet 4.5","configuration":"Claude Sonnet 4.5 · claude-sonnet-4-5-20250929","value":23.52,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-5"]},{"id":"claude-sonnet-4-6-thinking-auto-medium-effort","modelId":"model:anthropic/claude-sonnet-4-6","label":"Claude 4.6 Sonnet Thinking Medium Effort","configuration":"Claude 4.6 Sonnet Thinking Medium Effort · claude-sonnet-4-6-thinking-auto-medium-effort","value":63.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-sonnet-4-6"]},{"id":"claude-opus-4-7-xhigh-effort","modelId":"model:anthropic/claude-opus-4-7","label":"Claude 4.7 Opus Thinking xHigh Effort","configuration":"Claude 4.7 Opus Thinking xHigh Effort · claude-opus-4-7-xhigh-effort","value":59.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.anthropic.com/news/claude-opus-4-7"]},{"id":"deepseek-v3.2","modelId":"model:deepseek/deepseek-v3.2","label":"DeepSeek V3.2","configuration":"DeepSeek V3.2 · deepseek-v3.2","value":23.06,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/deepseek-ai/DeepSeek-V3.2"]},{"id":"devstral-2512","modelId":"model:mistral/devstral-2512","label":"Devstral 2","configuration":"Devstral 2 · devstral-2512","value":13.5,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mistral.ai/news/devstral-2-vibe-cli"]},{"id":"gemini-3-flash-preview-high","modelId":"model:google/gemini-3-flash-preview","label":"Gemini 3 Flash Preview High","configuration":"Gemini 3 Flash Preview High · gemini-3-flash-preview-high","value":74.86,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3-flash/"]},{"id":"gemini-3-pro-preview-11-2025-high","modelId":"model:google/gemini-3-pro-preview","label":"Gemini 3 Pro Preview High","configuration":"Gemini 3 Pro Preview High · gemini-3-pro-preview-11-2025-high","value":65.85,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/products/gemini/gemini-3/"]},{"id":"gemini-3.1-pro-preview-high","modelId":"model:google/gemini-3.1-pro-preview","label":"Gemini 3.1 Pro Preview High","configuration":"Gemini 3.1 Pro Preview High · gemini-3.1-pro-preview-high","value":79.1,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview/"]},{"id":"gemini-3.5-flash-high","modelId":"model:google/gemini-3.5-flash","label":"Gemini 3.5 Flash High","configuration":"Gemini 3.5 Flash High · gemini-3.5-flash-high","value":75.6,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/"]},{"id":"gemma-4-31b-it","modelId":"model:google/gemma-4-31b-it","label":"Gemma 4 31B","configuration":"Gemma 4 31B · gemma-4-31b-it","value":67.58,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/"]},{"id":"gpt-5-mini-high","modelId":"model:openai/gpt-5-mini","label":"GPT-5 Mini High","configuration":"GPT-5 Mini High · gpt-5-mini-high","value":65.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-nano-high","modelId":"model:openai/gpt-5-nano","label":"GPT-5 Nano High","configuration":"GPT-5 Nano High · gpt-5-nano-high","value":55.7,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5-pro-2025-10-06","modelId":"model:openai/gpt-5-pro","label":"GPT-5 Pro","configuration":"GPT-5 Pro · gpt-5-pro-2025-10-06","value":63.96,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5/"]},{"id":"gpt-5.1-2025-11-13-high","modelId":"model:openai/gpt-5.1","label":"GPT-5.1 High","configuration":"GPT-5.1 High · gpt-5.1-2025-11-13-high","value":63.9,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/gpt-5-1/"]},{"id":"gpt-5.1-codex","modelId":"model:openai/gpt-5.1-codex","label":"GPT-5.1 Codex","configuration":"GPT-5.1 Codex · gpt-5.1-codex","value":63.39,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex"]},{"id":"gpt-5.1-codex-max-high","modelId":"model:openai/gpt-5.1-codex-max","label":"GPT-5.1 Codex Max High","configuration":"GPT-5.1 Codex Max High · gpt-5.1-codex-max-high","value":70.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-max"]},{"id":"gpt-5.1-codex-mini","modelId":"model:openai/gpt-5.1-codex-mini","label":"GPT-5.1 Codex Mini","configuration":"GPT-5.1 Codex Mini · gpt-5.1-codex-mini","value":59.02,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.1-codex-mini"]},{"id":"gpt-5.2-2025-12-11-high","modelId":"model:openai/gpt-5.2","label":"GPT-5.2 High","configuration":"GPT-5.2 High · gpt-5.2-2025-12-11-high","value":61.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-5-2/"]},{"id":"gpt-5.2-codex","modelId":"model:openai/gpt-5.2-codex","label":"GPT-5.2 Codex","configuration":"GPT-5.2 Codex · gpt-5.2-codex","value":66.45,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.2-codex"]},{"id":"gpt-5.3-codex-high","modelId":"model:openai/gpt-5.3-codex","label":"GPT-5.3 Codex High","configuration":"GPT-5.3 Codex High · gpt-5.3-codex-high","value":65.38,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.3-codex"]},{"id":"gpt-5.4-xhigh","modelId":"model:openai/gpt-5.4","label":"GPT-5.4 Thinking xHigh Effort","configuration":"GPT-5.4 Thinking xHigh Effort · gpt-5.4-xhigh","value":70.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4"]},{"id":"gpt-5.4-mini-xhigh","modelId":"model:openai/gpt-5.4-mini","label":"GPT-5.4 Mini xHigh","configuration":"GPT-5.4 Mini xHigh · gpt-5.4-mini-xhigh","value":60.27,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-mini"]},{"id":"gpt-5.4-nano-xhigh","modelId":"model:openai/gpt-5.4-nano","label":"GPT-5.4 Nano xHigh","configuration":"GPT-5.4 Nano xHigh · gpt-5.4-nano-xhigh","value":67.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.4-nano"]},{"id":"gpt-5.5-xhigh","modelId":"model:openai/gpt-5.5","label":"GPT-5.5 Thinking xHigh Effort","configuration":"GPT-5.5 Thinking xHigh Effort · gpt-5.5-xhigh","value":73.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://platform.openai.com/docs/models/gpt-5.5"]},{"id":"gpt-oss-120b","modelId":"model:openai/gpt-oss-120b","label":"GPT OSS 120b","configuration":"GPT OSS 120b · gpt-oss-120b","value":50.29,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openai.com/index/introducing-gpt-oss/"]},{"id":"grok-4.3","modelId":"model:xai/grok-4.3","label":"Grok 4.3","configuration":"Grok 4.3 · grok-4.3","value":62.75,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://docs.x.ai/developers/models/grok-4.3"]},{"id":"kimi-k2-thinking","modelId":"model:moonshotai/kimi-k2-thinking","label":"Kimi K2 Thinking","configuration":"Kimi K2 Thinking · kimi-k2-thinking","value":62.03,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/moonshotai/Kimi-K2-Thinking"]},{"id":"kimi-k2.6-thinking","modelId":"model:moonshotai/kimi-k2.6","label":"Kimi K2.6 Thinking","configuration":"Kimi K2.6 Thinking · kimi-k2.6-thinking","value":64.36,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://www.kimi.com/blog/kimi-k2-6"]},{"id":"mimo-v2-pro","modelId":"model:xiaomi/mimo-v2-pro","label":"MiMo V2 Pro","configuration":"MiMo V2 Pro · mimo-v2-pro","value":43.22,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://mimo.xiaomi.com/mimo-v2-pro"]},{"id":"qwen3-235b-a22b-instruct-2507","modelId":"model:alibaba/qwen3-235b-a22b-instruct-2507","label":"Qwen 3 235B A22B Instruct 2507","configuration":"Qwen 3 235B A22B Instruct 2507 · qwen3-235b-a22b-instruct-2507","value":21.72,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://huggingface.co/Qwen/Qwen3-235B-A22B-Instruct-2507"]},{"id":"qwen3-30b-a3b-thinking","modelId":"model:alibaba/qwen3-30b-a3b","label":"Qwen 3 30B A3B","configuration":"Qwen 3 30B A3B · qwen3-30b-a3b-thinking","value":21.11,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-32b-thinking","modelId":"model:alibaba/qwen3-32b","label":"Qwen 3 32B","configuration":"Qwen 3 32B · qwen3-32b-thinking","value":17.77,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwenlm.github.io/blog/qwen3/"]},{"id":"qwen3-next-80b-a3b-instruct","modelId":"model:alibaba/qwen3-next-80b-a3b-instruct","label":"Qwen 3 Next 80B A3B Instruct","configuration":"Qwen 3 Next 80B A3B Instruct · qwen3-next-80b-a3b-instruct","value":19.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3-next-80b-a3b-thinking","modelId":"model:alibaba/qwen3-next-80b-a3b-thinking","label":"Qwen 3 Next 80B A3B Thinking","configuration":"Qwen 3 Next 80B A3B Thinking · qwen3-next-80b-a3b-thinking","value":41.54,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list"]},{"id":"qwen3.6-plus","modelId":"model:alibaba/qwen3.6-plus","label":"Qwen 3.6 Plus","configuration":"Qwen 3.6 Plus · qwen3.6-plus","value":58.34,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6/"]},{"id":"qwen3.6-27b","modelId":"model:alibaba/qwen3.6-27b","label":"Qwen 3.6 27B","configuration":"Qwen 3.6 27B · qwen3.6-27b","value":53.23,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.6-27b"]},{"id":"qwen3.6-flash","modelId":"model:alibaba/qwen3.6-flash","label":"Qwen 3.6 Flash","configuration":"Qwen 3.6 Flash · qwen3.6-flash","value":47.2,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://openrouter.ai/qwen/qwen3.6-flash/"]},{"id":"qwen3.7-max","modelId":"model:alibaba/qwen3.7-max","label":"Qwen 3.7 Max","configuration":"Qwen 3.7 Max · qwen3.7-max","value":74.04,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://qwen.ai/blog?id=qwen3.7"]},{"id":"nemotron-3-ultra-550b-a55b","modelId":"model:nvidia/nemotron-3-ultra-550b-a55b","label":"Nemotron 3 Ultra 550B A55B","configuration":"Nemotron 3 Ultra 550B A55B · nemotron-3-ultra-550b-a55b","value":58.19,"interval":null,"measuredAt":null,"evidenceUrls":["https://livebench.ai/table_2026_01_08.csv","https://build.nvidia.com/nvidia/nemotron-3-ultra-550b-a55b"]}]}]}