模型排行榜
每日同步
来源日期:2026-11-19
内容更新:2026-11-20 18:07(UTC+8)
图片生成网页
根据参考图片开发网页的独立榜单;不与文字网页开发分数混用。
| 排名 | 模型 | 机构 | 竞技场评分 | 置信区间 | 投票数 |
|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 (Max) | Anthropic | 1,758.4 | 1,740.3 – 1,776.5 | 1,421 |
| 2 | Claude Sonnet 5.5 (xHigh) | Anthropic | 1,746.4 | 1,733.0 – 1,759.9 | 2,630 |
| 3 | Claude Fable 5.1 (Max) | Anthropic | 1,722.7 | 1,711.3 – 1,734.0 | 3,741 |
| 4 | GPT-6 Astra (Max) | OpenAI | 1,718.6 | 1,707.7 – 1,729.4 | 4,253 |
| 5 | GPT-6.1 Sol (Max) | OpenAI | 1,694.5 | 1,682.1 – 1,706.9 | 3,043 |
| 6 | Qwen4-Max (xHigh) | 阿里巴巴 | 1,674.9 | 1,658.7 – 1,691.0 | 1,488 |
| 7 | GPT-6.1 Terra (Max) | OpenAI | 1,674.2 | 1,654.3 – 1,694.2 | 1,020 |
| 8 | Claude Opus 5 (Max) | Anthropic | 1,664.0 | 1,654.5 – 1,673.6 | 5,015 |
| 9 | GPT-6 Sol (Max) | OpenAI | 1,659.1 | 1,648.7 – 1,669.5 | 4,050 |
| 10 | Qwen3.8-Max (0902) | 阿里巴巴 | 1,638.6 | 1,619.6 – 1,657.7 | 1,079 |
| 11 | Muse Spark 1.3 (Max) | Meta | 1,635.8 | 1,627.3 – 1,644.2 | 6,850 |
| 12 | Claude Fable 5 (High) | Anthropic | 1,620.2 | 1,611.5 – 1,628.8 | 6,100 |
| 13 | Hy4 | 腾讯 | 1,618.6 | 1,603.0 – 1,634.3 | 1,495 |
| 14 | Qwen3.8-Max | 阿里巴巴 | 1,615.3 | 1,605.1 – 1,625.5 | 4,018 |
| 15 | DeepSeek-V4.1-Flash-20260910 (Max) | 深度求索 | 1,611.8 | 1,589.1 – 1,634.5 | 721 |
| 16 | Kimi K3.1 Pro (Max) | 月之暗面 | 1,608.8 | 1,568.7 – 1,649.0 | 224 |
| 17 | DeepSeek-V4.1-Flash-20261017 (Max) | 深度求索 | 1,608.1 | 1,598.3 – 1,617.9 | 4,670 |
| 18 | Grok 4.6 (High) | xai | 1,607.2 | 1,597.2 – 1,617.1 | 4,287 |
| 19 | GPT-5.6 Sol (xHigh, Codex Harness) | OpenAI | 1,606.0 | 1,596.9 – 1,615.1 | 5,363 |
| 20 | Gemini 3.7 Flash (High) | 谷歌 | 1,600.5 | 1,588.9 – 1,612.1 | 2,979 |
| 21 | Muse Spark 1.3 (xHigh) | Meta | 1,598.7 | 1,588.0 – 1,609.4 | 3,700 |
| 22 | MiMo-V2.6-Pro | 小米 | 1,595.5 | 1,585.1 – 1,605.9 | 3,991 |
| 23 | Gemini 3.8 Flash (High) | 谷歌 | 1,592.5 | 1,581.6 – 1,603.4 | 3,578 |
| 24 | MiMo-V2.6-Flash | 小米 | 1,592.1 | 1,581.4 – 1,602.8 | 3,642 |
| 25 | Grok 4.7 (xHigh) | xai | 1,591.5 | 1,578.3 – 1,604.8 | 2,253 |
| 26 | Kimi K3 (Max) | 月之暗面 | 1,586.4 | 1,577.1 – 1,595.8 | 4,969 |
| 27 | GLM-5.3-Flash | zai | 1,581.4 | 1,573.4 – 1,589.3 | 8,447 |
| 28 | Claude Opus 4.7 (High) | Anthropic | 1,579.4 | 1,570.1 – 1,588.8 | 6,737 |
| 29 | Qwen3.8-27B | 阿里巴巴 | 1,574.9 | 1,567.3 – 1,582.5 | 8,779 |
| 30 | GPT-6 Luna (Max) | OpenAI | 1,567.5 | 1,559.4 – 1,575.5 | 8,266 |
| 31 | Claude Opus 4.7 | Anthropic | 1,562.0 | 1,553.0 – 1,571.1 | 7,298 |
| 32 | Grok 4.5 | xai | 1,561.5 | 1,552.0 – 1,571.0 | 4,552 |
| 33 | Step 5 Preview (High) | 阶跃星辰 | 1,550.2 | 1,540.0 – 1,560.4 | 4,386 |
| 34 | Gemini 3.6 Flash (High) | 谷歌 | 1,544.4 | 1,529.8 – 1,559.0 | 1,741 |
| 35 | Claude Opus 4.6 (High) | Anthropic | 1,544.0 | 1,534.9 – 1,553.0 | 7,755 |
| 36 | GPT-5.6 Terra (xHigh, Codex Harness) | OpenAI | 1,539.5 | 1,530.3 – 1,548.6 | 5,177 |
| 37 | Claude Sonnet 5 (High) | Anthropic | 1,538.9 | 1,531.3 – 1,546.5 | 8,311 |
| 38 | Muse Spark 1.1 | Meta | 1,538.8 | 1,527.3 – 1,550.4 | 3,025 |
| 39 | Claude Opus 4.6 | Anthropic | 1,537.7 | 1,528.5 – 1,546.9 | 7,883 |
| 40 | GPT-5.5 (xHigh, Codex Harness) | OpenAI | 1,531.4 | 1,522.1 – 1,540.7 | 6,609 |
| 41 | Claude Sonnet 4.6 | Anthropic | 1,529.3 | 1,520.3 – 1,538.3 | 8,129 |
| 42 | Kimi K2.6 | 月之暗面 | 1,522.1 | 1,510.5 – 1,533.7 | 3,504 |
| 43 | GPT-5.5 (High, Codex Harness) | OpenAI | 1,519.1 | 1,510.0 – 1,528.1 | 7,273 |
| 44 | GPT-5.6 Luna (xHigh, Codex Harness) | OpenAI | 1,517.5 | 1,507.5 – 1,527.5 | 4,109 |
| 45 | Kimi K2.7 Code | 月之暗面 | 1,511.4 | 1,491.0 – 1,531.9 | 902 |
| 46 | Gemini 3.5 Flash (Medium) | 谷歌 | 1,505.0 | 1,493.0 – 1,516.9 | 2,757 |
| 47 | Seed2.1 Pro Preview | 字节跳动 | 1,504.8 | 1,495.6 – 1,514.1 | 4,687 |
| 48 | GPT-5.5 (Codex Harness) | OpenAI | 1,503.5 | 1,494.4 – 1,512.6 | 6,915 |
| 49 | Gemini 3.5 Flash (High) | 谷歌 | 1,500.0 | 1,486.0 – 1,514.1 | 1,959 |
| 50 | Gemini 3.1 Pro Preview | 谷歌 | 1,479.9 | 1,471.0 – 1,488.8 | 8,037 |
| 51 | MiniMax-M3 | MiniMax | 1,473.7 | 1,465.2 – 1,482.1 | 6,224 |
| 52 | Qwen3.6-Plus | 阿里巴巴 | 1,473.0 | 1,462.4 – 1,483.5 | 5,691 |
| 53 | Gemini 3 Flash | 谷歌 | 1,451.3 | 1,441.9 – 1,460.8 | 7,507 |
| 54 | Gemini 3.5 Flash-Lite | 谷歌 | 1,449.3 | 1,433.7 – 1,464.9 | 1,556 |
| 55 | GPT-5.3-Codex (Codex Harness) | OpenAI | 1,443.7 | 1,429.4 – 1,458.0 | 2,623 |
| 56 | Gemini 3 Pro | 谷歌 | 1,440.9 | 1,418.5 – 1,463.4 | 959 |
| 57 | GPT-5.4 | OpenAI | 1,440.8 | 1,429.2 – 1,452.5 | 3,349 |
| 58 | Kimi K2.5 (Thinking) | 月之暗面 | 1,439.2 | 1,421.7 – 1,456.7 | 1,641 |
| 59 | Qwen3.5-397B-A17B | 阿里巴巴 | 1,437.8 | 1,429.1 – 1,446.6 | 8,813 |
| 60 | Gemini 3 Flash (Minimal) | 谷歌 | 1,430.9 | 1,421.5 – 1,440.2 | 7,382 |
| 61 | GPT-5.1 (High) | OpenAI | 1,419.2 | 1,397.3 – 1,441.0 | 997 |
| 62 | Kimi K2.5 (Instant) | 月之暗面 | 1,410.7 | 1,388.4 – 1,432.9 | 966 |
| 63 | Inkling-Small | thinky | 1,409.9 | 1,390.3 – 1,429.5 | 998 |
| 64 | NVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4) | nvidia | 1,391.7 | 1,377.0 – 1,406.4 | 2,416 |
| 65 | Grok 4.3 | xai | 1,375.5 | 1,365.7 – 1,385.4 | 5,723 |
| 66 | Mistral Large 3 | mistral | 1,361.4 | 1,343.6 – 1,379.1 | 1,365 |
| 67 | GPT-5.1 | OpenAI | 1,348.3 | 1,326.7 – 1,369.8 | 1,119 |
| 68 | Gemini 3.1 Flash-Lite Preview | 谷歌 | 1,338.1 | 1,327.1 – 1,349.2 | 6,385 |
| 69 | Gemini 2.5 Pro | 谷歌 | 1,284.6 | 1,263.3 – 1,306.0 | 1,047 |
数据来源:Arena 官方图片生成网页。 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 18:07(UTC+8)。