大模型排行榜
Arena比较模型生成网页的效果,由参与者投票。
Arena AI 原榜 · 榜单日期
| 排名 | 模型 | 评分 | 对战票数 |
|---|---|---|---|
| 1 | Claude Fable 5.2 (xHigh)anthropic | 1,8231807–1839 | 1,993 |
| 2 | Claude Opus 5.5 (Max)anthropic | 1,8181807–1829 | 4,912 |
| 3 | Claude Sonnet 5.5 (xHigh)anthropic | 1,7711762–1781 | 6,721 |
| 4 | Gemini 4 Carbon (High)google | 1,7691753–1784 | 2,092 |
| 5 | Gemini 4 Neon (High)google | 1,7631749–1777 | 2,324 |
| 6 | GPT-6 Astra (Max)openai | 1,7631755–1771 | 9,316 |
| 7 | GPT-6.2 Astra (Max)openai | 1,7591744–1775 | 1,848 |
| 8 | GPT-6.2 Sol (Max)openai | 1,7551740–1769 | 1,947 |
| 9 | Qwen4-Max (xHigh)alibaba | 1,7501734–1765 | 1,808 |
| 10 | GPT-6.1 Sol (Max)openai | 1,7441734–1753 | 6,459 |
| 11 | Claude Fable 5.1 (Max)anthropic | 1,7411734–1749 | 10,591 |
| 12 | Muse Wonder Full (Max)meta | 1,7231709–1738 | 2,052 |
| 13 | Claude Sonnet 5.5 (High)anthropic | 1,7161708–1724 | 8,851 |
| 14 | Gemini 4 Hydrogen (High)google | 1,7011687–1715 | 2,039 |
| 15 | Claude Opus 5 (Max)anthropic | 1,6921686–1698 | 19,465 |
| 16 | GPT-6.1 Terra (Max)openai | 1,6841675–1694 | 5,170 |
| 17 | Kimi K3.1 Pro (Max)moonshot | 1,6831674–1692 | 5,958 |
| 18 | Gemini 4 Argon (High)google | 1,6811672–1690 | 5,978 |
| 19 | Qwen3.8-Max (0902)alibaba | 1,6751669–1681 | 25,208 |
| 20 | GPT-6 Sol (Max)openai | 1,6741666–1682 | 8,303 |
| 21 | Qwen3.8-Maxalibaba | 1,6711659–1683 | 3,356 |
| 22 | DeepSeek-V4.1-Flash-20261017 (Max)deepseek | 1,6591652–1665 | 17,765 |
| 23 | Kimi K3 (Max)moonshot | 1,6581653–1664 | 23,328 |
| 24 | Claude Opus 5 (High)anthropic | 1,6571651–1662 | 26,790 |
| 25 | Muse Spark 1.3 (Max)meta | 1,6541647–1660 | 13,332 |
| 26 | Hy4tencent | 1,6491637–1661 | 2,924 |
| 27 | Muse Wonder (Max)meta | 1,6481639–1656 | 6,477 |
| 28 | Inkling 2 Smallthinky | 1,6441626–1661 | 1,195 |
| 29 | Grok 4.7 (xHigh)xai | 1,6381630–1645 | 9,273 |
| 30 | MiMo-V2.6-Flashxiaomi | 1,6381630–1645 | 9,966 |
| 31 | Hy4 previewtencent | 1,6321622–1641 | 4,577 |
| 32 | Qwen3.8-Flash-Nextalibaba | 1,6261619–1633 | 12,404 |
| 33 | Claude Fable 5 (High)anthropic | 1,6261620–1632 | 16,874 |
| 34 | GLM-5.3 (Max)zai | 1,6221615–1628 | 13,737 |
| 35 | MiMo-V2.6-Proxiaomi | 1,6221615–1629 | 10,177 |
| 36 | DeepSeek-V4.1-Flash-20260910 (Max)deepseek | 1,6211610–1632 | 3,377 |
| 37 | Muse Spark 1.3 (xHigh)meta | 1,6211614–1628 | 12,861 |
| 38 | Grok 4.6 (High)xai | 1,6181611–1625 | 12,121 |
| 39 | GPT-5.6 Sol (xHigh, Codex Harness)openai | 1,6151610–1621 | 18,792 |
| 40 | GLM-5.3-Flashzai | 1,6101604–1616 | 24,714 |
| 41 | Muse Ember (Max)meta | 1,6091592–1626 | 1,322 |
| 42 | U2 Flashunisound | 1,6091592–1625 | 1,499 |
| 43 | GLM-5.2 (Max)zai | 1,6071602–1613 | 17,837 |
| 44 | Claude Haiku 5.5 (High)anthropic | 1,5961588–1604 | 8,314 |
| 45 | Qwen3.8-27Balibaba | 1,5891583–1595 | 22,065 |
| 46 | Gemini 3.7 Flash (High)google | 1,5891582–1595 | 14,496 |
| 47 | Gemini 3.8 Flash (High)google | 1,5841577–1590 | 15,072 |
| 48 | DeepSeek-V4-Flash (High)deepseek | 1,5821572–1591 | 4,809 |
| 49 | MiniMax-M3.1-Flashminimax | 1,5821573–1591 | 5,948 |
| 50 | DeepSeek-V4-Pro (High, 20260813)deepseek | 1,5811571–1590 | 4,593 |