模型排行榜
每日同步
来源日期:2026-11-09
内容更新:2026-11-10 18:05(UTC+8)
文档分析
文档分析综合榜,使用官方公开原始评分;竞技场评分越高越好。
| 排名 | 模型 | 机构 | 竞技场评分 | 置信区间 | 投票数 |
|---|---|---|---|---|---|
| 1 | Claude Fable 5.1 (Max) | Anthropic | 1,514.8 | 1,499.7 – 1,530.0 | 1,453 |
| 2 | Claude Opus 5 (High) | Anthropic | 1,508.7 | 1,503.0 – 1,514.3 | 16,189 |
| 3 | Claude Opus 4.6 (High) | Anthropic | 1,507.6 | 1,501.2 – 1,514.0 | 29,456 |
| 4 | Claude Opus 4.6 | Anthropic | 1,506.2 | 1,501.0 – 1,511.4 | 47,264 |
| 5 | Claude Fable 5 (High) | Anthropic | 1,496.6 | 1,490.0 – 1,503.3 | 12,357 |
| 6 | Claude Opus 4.7 (High) | Anthropic | 1,495.7 | 1,489.3 – 1,502.1 | 23,327 |
| 7 | Claude Opus 4.7 | Anthropic | 1,494.8 | 1,488.4 – 1,501.2 | 23,628 |
| 8 | GPT-5.5 (High) | OpenAI | 1,484.3 | 1,478.9 – 1,489.6 | 27,324 |
| 9 | Gemini 3.8 Flash (High) | 谷歌 | 1,481.9 | 1,474.0 – 1,489.9 | 6,950 |
| 10 | Claude Sonnet 4.6 | Anthropic | 1,481.9 | 1,476.0 – 1,487.7 | 61,055 |
| 11 | GPT-5.5 | OpenAI | 1,480.9 | 1,475.6 – 1,486.2 | 27,751 |
| 12 | GPT-5.6 Sol (xHigh) | OpenAI | 1,480.0 | 1,473.2 – 1,486.8 | 9,649 |
| 13 | Gemini 3.7 Flash (High) | 谷歌 | 1,476.5 | 1,468.9 – 1,484.1 | 7,680 |
| 14 | Claude Opus 4.8 (High) | Anthropic | 1,474.6 | 1,467.7 – 1,481.4 | 12,069 |
| 15 | GPT-5.4 | OpenAI | 1,472.9 | 1,467.0 – 1,478.9 | 35,183 |
| 16 | Muse Spark 1.3 (Max) | Meta | 1,472.9 | 1,465.1 – 1,480.7 | 7,205 |
| 17 | GPT-5.6 Terra (xHigh) | OpenAI | 1,467.7 | 1,461.2 – 1,474.1 | 11,412 |
| 18 | GPT-6 Astra (Max) | OpenAI | 1,467.3 | 1,458.7 – 1,475.8 | 5,723 |
| 19 | Claude Opus 4.5 (20251101) | Anthropic | 1,465.7 | 1,455.5 – 1,475.9 | 8,304 |
| 20 | Claude Opus 4.8 | Anthropic | 1,465.1 | 1,458.7 – 1,471.4 | 14,729 |
| 21 | Muse Spark 1.1 | Meta | 1,463.5 | 1,456.4 – 1,470.6 | 8,934 |
| 22 | Gemini 3.5 Flash (High) | 谷歌 | 1,461.8 | 1,452.6 – 1,470.9 | 4,248 |
| 23 | Gemini 3.5 Flash (Medium) | 谷歌 | 1,460.6 | 1,452.9 – 1,468.4 | 6,835 |
| 24 | Claude Sonnet 5 (High) | Anthropic | 1,459.5 | 1,453.0 – 1,466.1 | 11,044 |
| 25 | Gemini 3.6 Flash (High) | 谷歌 | 1,456.1 | 1,445.4 – 1,466.8 | 3,031 |
| 26 | GPT-5.6 Luna (xHigh) | OpenAI | 1,453.8 | 1,446.8 – 1,460.8 | 8,832 |
| 27 | Grok 4.5 | xai | 1,451.5 | 1,443.9 – 1,459.1 | 7,447 |
| 28 | Kimi K2.6 | 月之暗面 | 1,450.4 | 1,442.9 – 1,458.0 | 11,809 |
| 29 | Claude Sonnet 4.5 (20250929) | Anthropic | 1,449.3 | 1,443.3 – 1,455.4 | 32,916 |
| 30 | Muse Spark | Meta | 1,443.6 | 1,425.4 – 1,461.7 | 1,116 |
| 31 | Gemini 3.1 Pro Preview | 谷歌 | 1,443.3 | 1,438.1 – 1,448.4 | 52,244 |
| 32 | Grok 4.6 (High) | xai | 1,438.1 | 1,430.3 – 1,445.8 | 7,669 |
| 33 | Qwen3.7-Plus | 阿里巴巴 | 1,437.4 | 1,430.0 – 1,444.9 | 7,426 |
| 34 | MiniMax-M3 | MiniMax | 1,436.9 | 1,428.8 – 1,445.0 | 6,617 |
| 35 | Gemma 4 31B | 谷歌 | 1,434.3 | 1,427.6 – 1,441.1 | 15,431 |
| 36 | Gemini 3 Pro | 谷歌 | 1,433.9 | 1,425.1 – 1,442.6 | 11,338 |
| 37 | Kimi K2.5 (Thinking) | 月之暗面 | 1,430.7 | 1,424.0 – 1,437.4 | 22,601 |
| 38 | Gemini 2.5 Pro | 谷歌 | 1,422.8 | 1,416.7 – 1,428.9 | 26,635 |
| 39 | Claude Haiku 4.5 (20251001) | Anthropic | 1,420.4 | 1,415.1 – 1,425.7 | 39,423 |
| 40 | GLM-5V-Turbo | zai | 1,418.8 | 1,411.4 – 1,426.2 | 9,136 |
| 41 | Grok 4.20 Beta (Reasoning, 0309) | xai | 1,417.1 | 1,410.2 – 1,424.0 | 21,926 |
| 42 | Gemini 3 Flash | 谷歌 | 1,414.3 | 1,405.0 – 1,423.6 | 7,538 |
| 43 | GPT-5.2 (High) | OpenAI | 1,403.9 | 1,394.7 – 1,413.1 | 7,552 |
| 44 | GPT-5.5 Instant | OpenAI | 1,403.2 | 1,395.4 – 1,411.1 | 8,834 |
| 45 | GPT-5.1 | OpenAI | 1,403.0 | 1,393.9 – 1,412.1 | 8,736 |
| 46 | GPT-5.2 | OpenAI | 1,402.1 | 1,396.3 – 1,407.9 | 29,726 |
数据来源:Arena 官方文档分析。 使用官方公开数据集,按 CC BY 4.0 标注来源。 各榜单独同步;来源未发布新数据时,内容日期保持不变。
竞技场评分与任务、评测方式有关,不同子榜的分数不宜直接比较。 最近检查:2026-11-22 18:04(UTC+8)。