大模型排行榜
Arena评测同步于
评分说明与数据来源
综合榜采用 Epoch 能力指数(ECI)。各分类使用页面标注的单项评测,展示该评测的原始百分比分,不合并不同榜单的分数。
评测有多个运行记录时,优先取日期最近的一次。模型详情列出实际参测配置。综合榜展示评测覆盖项数和 ECI 的 90% 可信区间。
考查模型根据行为描述重建开源代码库的能力。
MirrorCode ↗编程榜
41 个模型
| 01 | Claude Sonnet 5.5Anthropic · 2026-09-28 | 97.78% |
|---|---|---|
| 02 | Claude Opus 5.5Anthropic · 2026-09-22 | 95.56% |
| 03 | Claude Opus 5Anthropic · 2026-07-24 | 87.78% |
| 04 | Claude Haiku 5.5Anthropic · 2026-10-07 | 78.89% |
| 05 | GPT-6.1 TerraOpenAI · 2026-10-26 | 76.67% |
| 06 | Claude Fable 5.1Anthropic · 2026-09-01 | 73.3% |
| 07 | Claude Fable 5Anthropic · 2026-06-09 | 70.0% |
| 08 | GPT-6.2 AstraOpenAI · 2026-11-17 | 67.78% |
| 09 | Claude Opus 4.8Anthropic · 2026-05-28 | 65.56% |
| 10 | DeepSeek V4.1 FlashDeepSeek · 2026-09-09 · 开放权重 | 64.44% |
| 11 | Claude Sonnet 5Anthropic · 2026-06-30 | 63.33% |
| 12 | GPT-6.1 SolOpenAI · 2026-09-29 | 61.11% |
| 13 | GPT-6 AstraOpenAI · 2026-09-03 | 60.0% |
| 14 | Gemini 4 ArgonGoogle DeepMind · 2026-10-13 | 58.89% |
| 15 | Grok 4.7xAI · 2026-09-21 | 54.44% |
| 15 | Kimi K3Moonshot · 2026-07-16 · 开放权重 | 54.44% |
| 15 | Muse Spark 1.3Meta AI · 2026-09-02 | 54.44% |
| 18 | Qwen 3.8 MaxAlibaba · 2026-08-02 | 52.22% |
| 19 | GPT-6 SolOpenAI · 2026-09-22 | 48.89% |
| 20 | Qwen3.8 Max (0902)Alibaba · 2026-09-01 | 47.78% |
| 21 | Muse WonderMeta AI · 2026-10-12 | 45.56% |
| 22 | Claude Opus 4.7Anthropic · 2026-04-16 | 43.33% |
| 22 | Kimi K3.1 ProMoonshot · 2026-10-23 · 开放权重 | 43.33% |
| 24 | Gemini 3.7 FlashGoogle DeepMind · 2026-08-13 | 41.11% |
| 24 | GPT-5.6 TerraOpenAI · 2026-07-09 | 41.11% |
| 24 | Grok 4.6xAI · 2026-08-12 | 41.11% |
| 27 | Mistral Large 4 PreviewMistral AI · 2026-10-06 · 开放权重 | 40.0% |
| 28 | Gemini 3.8 FlashGoogle DeepMind · 2026-09-02 | 38.89% |
| 29 | GLM-5.2Z.ai (Zhipu AI) · 2026-06-16 · 开放权重 | 36.67% |
| 30 | Claude Opus 4.6Anthropic · 2026-02-05 | 33.33% |
| 30 | DeepSeek V4 Pro 0813DeepSeek · 2026-08-13 · 开放权重 | 33.33% |
| 32 | GPT-5.2OpenAI · 2025-12-11 | 32.22% |
| 32 | Grok 4.5xAI · 2026-07-08 | 32.22% |
| 34 | GPT-6 LunaOpenAI · 2026-09-22 | 26.67% |
| 35 | Gemini 3.6 FlashGoogle DeepMind · 2026-07-21 | 23.33% |
| 35 | GPT-5.4OpenAI · 2026-03-05 | 23.33% |
| 35 | GPT-5.5OpenAI · 2026-04-23 | 23.33% |
| 38 | GPT-5.6 LunaOpenAI · 2026-07-09 | 21.11% |
| 39 | GPT-5.6 SolOpenAI · 2026-07-09 | 20.0% |
| 40 | Gemini 3.1 ProGoogle DeepMind · 2026-02-19 | 8.89% |
| 41 | GPT-5OpenAI · 2025-08-07 | 4.44% |