模型排行榜
每日同步
来源日期:2026-11-20
内容更新:2026-11-21 18:06(UTC+8)
网页开发
根据文字需求开发网页的综合榜;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 5.2 (xHigh) | Anthropic | 1,823.6 | 1,807.8 – 1,839.5 | 2,042 |
| 2 | Claude Opus 5.5 (Max) | Anthropic | 1,819.0 | 1,808.1 – 1,830.0 | 4,952 |
| 3 | Claude Sonnet 5.5 (xHigh) | Anthropic | 1,772.0 | 1,762.7 – 1,781.2 | 6,785 |
| 4 | Gemini 4 Carbon (High) | 谷歌 | 1,768.7 | 1,753.3 – 1,784.1 | 2,093 |
| 5 | GPT-6.2 Astra (Max) | OpenAI | 1,766.2 | 1,750.9 – 1,781.5 | 1,913 |
| 6 | Gemini 4 Neon (High) | 谷歌 | 1,764.9 | 1,751.2 – 1,778.6 | 2,305 |
| 7 | GPT-6 Astra (Max) | OpenAI | 1,764.0 | 1,755.8 – 1,772.1 | 9,391 |
| 8 | GPT-6.2 Sol (Max) | OpenAI | 1,754.0 | 1,739.5 – 1,768.6 | 2,067 |
| 9 | Qwen4-Max (xHigh) | 阿里巴巴 | 1,750.9 | 1,735.4 – 1,766.4 | 1,797 |
| 10 | GPT-6.1 Sol (Max) | OpenAI | 1,743.4 | 1,734.2 – 1,752.5 | 6,538 |
| 11 | Claude Fable 5.1 (Max) | Anthropic | 1,741.9 | 1,734.3 – 1,749.5 | 10,654 |
| 12 | Muse Wonder Full (Max) | Meta | 1,722.6 | 1,708.2 – 1,737.1 | 2,056 |
| 13 | Claude Sonnet 5.5 (High) | Anthropic | 1,715.5 | 1,707.6 – 1,723.4 | 8,933 |
| 14 | Gemini 4 Hydrogen (High) | 谷歌 | 1,700.6 | 1,686.4 – 1,714.9 | 2,021 |
| 15 | Claude Opus 5 (Max) | Anthropic | 1,691.6 | 1,685.5 – 1,697.7 | 19,462 |
| 16 | GPT-6.1 Terra (Max) | OpenAI | 1,685.4 | 1,675.9 – 1,694.9 | 5,239 |
| 17 | Kimi K3.1 Pro (Max) | 月之暗面 | 1,683.0 | 1,674.0 – 1,691.9 | 6,022 |
| 18 | Gemini 4 Argon (High) | 谷歌 | 1,682.4 | 1,673.5 – 1,691.4 | 6,016 |
| 19 | GPT-6 Sol (Max) | OpenAI | 1,674.4 | 1,666.6 – 1,682.3 | 8,376 |
| 20 | Qwen3.8-Max (0902) | 阿里巴巴 | 1,673.6 | 1,667.8 – 1,679.5 | 25,449 |
| 21 | Qwen3.8-Max | 阿里巴巴 | 1,670.4 | 1,658.2 – 1,682.5 | 3,353 |
| 22 | Kimi K3 (Max) | 月之暗面 | 1,659.2 | 1,653.6 – 1,664.7 | 23,391 |
| 23 | DeepSeek-V4.1-Flash-20261017 (Max) | 深度求索 | 1,658.0 | 1,651.6 – 1,664.3 | 18,037 |
| 24 | Claude Opus 5 (High) | Anthropic | 1,656.5 | 1,651.2 – 1,661.8 | 26,785 |
| 25 | Muse Spark 1.3 (Max) | Meta | 1,653.0 | 1,646.3 – 1,659.7 | 13,410 |
| 26 | Hy4 | 腾讯 | 1,650.5 | 1,638.5 – 1,662.5 | 2,909 |
| 27 | Muse Wonder (Max) | Meta | 1,647.0 | 1,638.5 – 1,655.5 | 6,570 |
| 28 | Inkling 2 Small | thinky | 1,642.7 | 1,624.9 – 1,660.4 | 1,204 |
| 29 | Grok 4.7 (xHigh) | xai | 1,638.9 | 1,631.4 – 1,646.4 | 9,336 |
| 30 | MiMo-V2.6-Flash | 小米 | 1,637.5 | 1,630.1 – 1,644.9 | 10,015 |
| 31 | Hy4 preview | 腾讯 | 1,631.4 | 1,621.7 – 1,641.2 | 4,575 |
| 32 | Claude Fable 5 (High) | Anthropic | 1,625.9 | 1,619.9 – 1,631.9 | 16,876 |
| 33 | Qwen3.8-Flash-Next | 阿里巴巴 | 1,625.7 | 1,618.9 – 1,632.5 | 12,473 |
| 34 | GLM-5.3 (Max) | zai | 1,622.7 | 1,616.0 – 1,629.3 | 13,776 |
| 35 | MiMo-V2.6-Pro | 小米 | 1,622.1 | 1,614.8 – 1,629.3 | 10,272 |
| 36 | DeepSeek-V4.1-Flash-20260910 (Max) | 深度求索 | 1,621.3 | 1,610.2 – 1,632.4 | 3,374 |
| 37 | Muse Spark 1.3 (xHigh) | Meta | 1,620.8 | 1,614.1 – 1,627.5 | 12,932 |
| 38 | Grok 4.6 (High) | xai | 1,618.0 | 1,611.0 – 1,624.9 | 12,125 |
| 39 | GPT-5.6 Sol (xHigh, Codex Harness) | OpenAI | 1,615.4 | 1,609.7 – 1,621.1 | 18,779 |
| 40 | GLM-5.3-Flash | zai | 1,610.5 | 1,604.4 – 1,616.5 | 24,789 |
| 41 | U2 Flash | unisound | 1,609.4 | 1,592.9 – 1,625.8 | 1,494 |
| 42 | Muse Ember (Max) | Meta | 1,608.3 | 1,591.4 – 1,625.2 | 1,316 |
| 43 | GLM-5.2 (Max) | zai | 1,607.5 | 1,601.6 – 1,613.4 | 17,837 |
| 44 | Claude Haiku 5.5 (High) | Anthropic | 1,595.4 | 1,587.6 – 1,603.1 | 8,448 |
| 45 | Qwen3.8-27B | 阿里巴巴 | 1,589.5 | 1,583.6 – 1,595.4 | 22,159 |
| 46 | Gemini 3.7 Flash (High) | 谷歌 | 1,588.7 | 1,582.2 – 1,595.2 | 14,501 |
| 47 | Gemini 3.8 Flash (High) | 谷歌 | 1,584.2 | 1,577.8 – 1,590.6 | 15,142 |
| 48 | DeepSeek-V4-Flash (High) | 深度求索 | 1,582.1 | 1,572.7 – 1,591.5 | 4,803 |
| 49 | MiniMax-M3.1-Flash | MiniMax | 1,581.4 | 1,572.6 – 1,590.3 | 5,996 |
| 50 | DeepSeek-V4-Pro (High, 20260813) | 深度求索 | 1,580.7 | 1,571.1 – 1,590.2 | 4,588 |
| 51 | GPT-6 Luna (Max) | OpenAI | 1,578.3 | 1,572.5 – 1,584.2 | 32,346 |
| 52 | Mistral Large 4 | mistral | 1,573.2 | 1,560.7 – 1,585.7 | 2,706 |
| 53 | Step 5 Preview (High) | 阶跃星辰 | 1,570.1 | 1,562.3 – 1,578.0 | 8,328 |
| 54 | Claude Opus 4.7 (High) | Anthropic | 1,557.4 | 1,551.6 – 1,563.2 | 17,558 |
| 55 | Claude Opus 4.7 | Anthropic | 1,555.5 | 1,549.9 – 1,561.2 | 17,870 |
| 56 | Claude Opus 4.8 (High) | Anthropic | 1,555.5 | 1,549.7 – 1,561.3 | 17,949 |
| 57 | Grok 4.5 | xai | 1,552.4 | 1,545.7 – 1,559.1 | 10,754 |
| 58 | Claude Opus 4.6 (High) | Anthropic | 1,546.4 | 1,540.8 – 1,551.9 | 19,013 |
| 59 | Muse Spark 1.1 | Meta | 1,541.3 | 1,533.6 – 1,548.9 | 7,805 |
| 60 | Claude Sonnet 5 (High) | Anthropic | 1,539.4 | 1,533.6 – 1,545.2 | 16,693 |
| 61 | Gemini 3.6 Flash (High) | 谷歌 | 1,537.8 | 1,530.2 – 1,545.4 | 8,464 |
| 62 | Claude Opus 4.6 | Anthropic | 1,536.7 | 1,531.3 – 1,542.1 | 20,340 |
| 63 | Claude Opus 4.8 | Anthropic | 1,536.5 | 1,530.8 – 1,542.2 | 16,886 |
| 64 | Muse Spark 1.2 (xHigh) | Meta | 1,533.3 | 1,519.5 – 1,547.1 | 2,112 |
| 65 | Muse Spark Open (Max) | Meta | 1,527.3 | 1,513.1 – 1,541.6 | 2,015 |
| 66 | Claude Sonnet 4.6 | Anthropic | 1,521.9 | 1,516.7 – 1,527.1 | 22,529 |
| 67 | GPT-5.6 Terra (xHigh, Codex Harness) | OpenAI | 1,520.4 | 1,514.3 – 1,526.5 | 15,387 |
| 68 | GPT-5.6 Luna (xHigh, Codex Harness) | OpenAI | 1,519.1 | 1,512.6 – 1,525.5 | 12,139 |
| 69 | Seed2.1 Pro Preview | 字节跳动 | 1,518.3 | 1,512.8 – 1,523.8 | 18,649 |
| 70 | Qwen3.7-Max (20260517) | 阿里巴巴 | 1,515.7 | 1,508.3 – 1,523.1 | 8,713 |
| 71 | GPT-5.5 (xHigh, Codex Harness) | OpenAI | 1,512.4 | 1,506.4 – 1,518.4 | 15,348 |
| 72 | Kimi K2.6 | 月之暗面 | 1,509.1 | 1,501.9 – 1,516.4 | 9,998 |
| 73 | GLM-5.1 | zai | 1,508.1 | 1,501.2 – 1,514.9 | 10,506 |
| 74 | Hy3 | 腾讯 | 1,505.8 | 1,498.1 – 1,513.5 | 7,638 |
| 75 | Gemini 3.5 Flash (High) | 谷歌 | 1,499.8 | 1,492.4 – 1,507.1 | 9,122 |
| 76 | Claude Opus 4.5 (High, 32K, 20251101) | Anthropic | 1,493.6 | 1,485.3 – 1,502.0 | 10,961 |
| 77 | Gemini 3.5 Flash (Medium) | 谷歌 | 1,490.1 | 1,483.4 – 1,496.8 | 11,842 |
| 78 | GPT-5.5 (High, Codex Harness) | OpenAI | 1,487.3 | 1,481.4 – 1,493.3 | 17,581 |
| 79 | Spark-X2.6 | synlan | 1,485.7 | 1,466.5 – 1,505.0 | 1,175 |
| 80 | MiniMax-M3 | MiniMax | 1,484.4 | 1,478.9 – 1,489.8 | 20,962 |
| 81 | Qwen3.6-Max-Preview | 阿里巴巴 | 1,481.3 | 1,468.8 – 1,493.9 | 2,685 |
| 82 | MiMo-V2.5-Pro | 小米 | 1,479.3 | 1,474.1 – 1,484.6 | 21,961 |
| 83 | Kimi K2.7 Code | 月之暗面 | 1,474.9 | 1,465.4 – 1,484.4 | 4,869 |
| 84 | Claude Opus 4.5 (20251101) | Anthropic | 1,468.9 | 1,461.7 – 1,476.2 | 13,332 |
| 85 | DeepSeek-V4-Pro Preview (High) | 深度求索 | 1,463.8 | 1,457.5 – 1,470.2 | 12,908 |
| 86 | Qwen3.6-Plus | 阿里巴巴 | 1,461.9 | 1,456.3 – 1,467.5 | 18,217 |
| 87 | GPT-5.4 (High, Codex Harness) | OpenAI | 1,461.5 | 1,442.6 – 1,480.5 | 1,336 |
| 88 | GPT-5.5 (Codex Harness) | OpenAI | 1,455.7 | 1,449.8 – 1,461.7 | 15,555 |
| 89 | DeepSeek-V4-Pro | 深度求索 | 1,448.4 | 1,442.1 – 1,454.8 | 13,738 |
| 90 | Gemini 3.1 Pro Preview | 谷歌 | 1,446.3 | 1,441.2 – 1,451.4 | 23,808 |
| 91 | GPT-5.4 (Medium, Codex Harness) | OpenAI | 1,441.7 | 1,423.4 – 1,460.1 | 1,364 |
| 92 | Gemini 3 Pro | 谷歌 | 1,441.1 | 1,432.6 – 1,449.6 | 13,968 |
| 93 | Gemini 3 Flash | 谷歌 | 1,440.1 | 1,431.5 – 1,448.8 | 10,801 |
| 94 | MiMo-V2.5 | 小米 | 1,437.9 | 1,431.8 – 1,443.9 | 14,588 |
| 95 | Beam 1.0 501B A23B | reflection | 1,436.8 | 1,423.1 – 1,450.5 | 2,494 |
| 96 | Kimi K2.5 (Thinking) | 月之暗面 | 1,436.4 | 1,431.1 – 1,441.6 | 19,769 |
| 97 | GLM-4.7 | zai | 1,435.9 | 1,423.8 – 1,447.9 | 3,992 |
| 98 | GLM-5 | zai | 1,434.9 | 1,426.7 – 1,443.1 | 7,430 |
| 99 | MiMo-V2-Pro | 小米 | 1,434.8 | 1,426.6 – 1,443.0 | 7,176 |
| 100 | Gemini 3.5 Flash-Lite | 谷歌 | 1,433.4 | 1,392.0 – 1,474.7 | 230 |
| 101 | DeepSeek-V4-Flash Preview (High) | 深度求索 | 1,431.1 | 1,424.1 – 1,438.1 | 9,960 |
| 102 | GPT-5.2 | OpenAI | 1,419.0 | 1,396.2 – 1,441.7 | 1,074 |
| 103 | GPT-5 (Medium) | OpenAI | 1,417.5 | 1,401.9 – 1,433.2 | 3,075 |
| 104 | Inkling | thinky | 1,414.1 | 1,408.1 – 1,420.0 | 17,085 |
| 105 | Inkling-Small | thinky | 1,407.7 | 1,398.5 – 1,416.9 | 5,197 |
| 106 | GPT-5.3-Codex (Codex Harness) | OpenAI | 1,406.1 | 1,392.4 – 1,419.8 | 2,631 |
| 107 | Kimi K2.5 (Instant) | 月之暗面 | 1,405.6 | 1,393.5 – 1,417.7 | 3,242 |
| 108 | NVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4) | nvidia | 1,400.7 | 1,390.7 – 1,410.6 | 6,264 |
| 109 | GPT-5.4 | OpenAI | 1,400.0 | 1,390.1 – 1,409.9 | 4,626 |
| 110 | Qwen3.5-397B-A17B | 阿里巴巴 | 1,399.9 | 1,394.8 – 1,405.1 | 20,811 |
| 111 | GPT-5.4 mini (High) | OpenAI | 1,398.2 | 1,391.6 – 1,404.8 | 11,782 |
| 112 | MiniMax-M2.7 | MiniMax | 1,398.0 | 1,392.2 – 1,403.9 | 15,682 |
| 113 | GPT-5.1 (Medium) | OpenAI | 1,397.2 | 1,385.6 – 1,408.9 | 4,878 |
| 114 | Claude Sonnet 4.5 (High, 32K, 20250929) | Anthropic | 1,393.9 | 1,386.2 – 1,401.7 | 13,162 |
| 115 | Claude Opus 4.1 (20250805) | Anthropic | 1,390.4 | 1,379.6 – 1,401.2 | 6,881 |
| 116 | MiniMax-M2.5 | MiniMax | 1,385.9 | 1,377.5 – 1,394.3 | 7,580 |
| 117 | Claude Sonnet 4.5 (20250929) | Anthropic | 1,384.8 | 1,377.5 – 1,392.2 | 15,616 |
| 118 | Gemini 3 Flash (Minimal) | 谷歌 | 1,384.5 | 1,379.2 – 1,389.7 | 21,951 |
| 119 | MiniMax-M2.1 Preview | MiniMax | 1,384.2 | 1,374.6 – 1,393.8 | 7,598 |
| 120 | Grok 4.20 Beta (Reasoning, 0309) | xai | 1,376.4 | 1,370.3 – 1,382.5 | 15,240 |
| 121 | GPT-5.3-Codex (Codex Harness) | OpenAI | 1,370.8 | 1,359.4 – 1,382.1 | 3,637 |
| 122 | Gemma 4 31B | 谷歌 | 1,370.2 | 1,364.3 – 1,376.2 | 16,567 |
| 123 | Solar Pro 4 | upstage | 1,369.2 | 1,358.7 – 1,379.7 | 4,481 |
| 124 | DeepSeek-V3.2 (Thinking) | 深度求索 | 1,362.5 | 1,353.3 – 1,371.6 | 6,872 |
| 125 | Qwen3.5-122B-A10B | 阿里巴巴 | 1,360.2 | 1,352.5 – 1,367.9 | 8,194 |
| 126 | Hy3 preview | 腾讯 | 1,358.8 | 1,341.3 – 1,376.3 | 1,425 |
| 127 | Gemma 4 26B A4B | 谷歌 | 1,357.6 | 1,340.6 – 1,374.6 | 1,487 |
| 128 | Grok 4.3 | xai | 1,357.5 | 1,351.2 – 1,363.8 | 14,661 |
| 129 | Qwen3.5-27B | 阿里巴巴 | 1,357.0 | 1,348.9 – 1,365.1 | 7,746 |
| 130 | Muse Glimmer | Meta | 1,353.0 | 1,337.3 – 1,368.7 | 1,622 |
| 131 | Laguna M.1 | poolside | 1,350.1 | 1,340.7 – 1,359.4 | 5,298 |
| 132 | GPT-5.1 | OpenAI | 1,342.3 | 1,333.8 – 1,350.9 | 10,628 |
| 133 | GLM-4.6 | zai | 1,342.0 | 1,331.1 – 1,352.8 | 6,718 |
| 134 | GPT-5.2-Codex | OpenAI | 1,338.7 | 1,329.7 – 1,347.7 | 6,670 |
| 135 | GPT-5.1-Codex | OpenAI | 1,338.0 | 1,326.4 – 1,349.7 | 4,968 |
| 136 | MiMo-V2-Flash (Non-Thinking) | 小米 | 1,331.6 | 1,321.8 – 1,341.4 | 5,649 |
| 137 | Claude Haiku 4.5 (20251001) | Anthropic | 1,330.2 | 1,325.2 – 1,335.2 | 28,713 |
| 138 | DeepSeek-V3.2 | 深度求索 | 1,325.7 | 1,317.7 – 1,333.8 | 9,319 |
| 139 | Kimi K2 Thinking Turbo | 月之暗面 | 1,321.8 | 1,314.3 – 1,329.3 | 12,997 |
| 140 | Laguna XS.2 | poolside | 1,304.0 | 1,293.2 – 1,314.7 | 4,156 |
| 141 | MiniMax-M2 | MiniMax | 1,298.9 | 1,287.9 – 1,310.0 | 6,658 |
| 142 | MiMo-V2-Flash (Thinking) | 小米 | 1,291.0 | 1,274.5 – 1,307.5 | 1,666 |
| 143 | Qwen3-Coder-480B-A35B-Instruct | 阿里巴巴 | 1,274.4 | 1,266.5 – 1,282.3 | 12,631 |
| 144 | DeepSeek-V3.2-Exp | 深度求索 | 1,271.4 | 1,257.8 – 1,285.0 | 3,935 |
| 145 | Mistral Medium 3.5 | mistral | 1,263.4 | 1,248.5 – 1,278.3 | 2,275 |
| 146 | KAT-Coder-Pro V1 | 1,257.4 | 1,237.8 – 1,276.9 | 1,494 | |
| 147 | Gemini 3.1 Flash-Lite Preview | 谷歌 | 1,255.8 | 1,248.9 – 1,262.7 | 14,234 |
| 148 | Qwen3.5-35B-A3B | 阿里巴巴 | 1,253.4 | 1,235.9 – 1,270.8 | 1,610 |
| 149 | GPT-5.1-Codex-Mini | OpenAI | 1,243.5 | 1,221.2 – 1,265.8 | 1,135 |
| 150 | Qwen3.5-Flash | 阿里巴巴 | 1,243.5 | 1,223.8 – 1,263.2 | 1,321 |
| 151 | Grok 4.1 Fast (Reasoning) | xai | 1,241.7 | 1,230.5 – 1,252.9 | 5,452 |
| 152 | Trinity Large Thinking | 1,236.6 | 1,216.5 – 1,256.7 | 1,345 | |
| 153 | Mistral Large 3 | mistral | 1,232.6 | 1,206.8 – 1,258.4 | 822 |
| 154 | Gemini 2.5 Pro | 谷歌 | 1,226.6 | 1,210.7 – 1,242.5 | 2,705 |
| 155 | Grok 4.1 (Thinking) | xai | 1,216.0 | 1,191.0 – 1,241.0 | 955 |
| 156 | Devstral 2 | mistral | 1,195.5 | 1,174.7 – 1,216.3 | 1,176 |
| 157 | Granite 4.1 8B | ibm | 1,192.8 | 1,174.3 – 1,211.3 | 1,759 |
| 158 | Mercury 2 | inception-ai | 1,170.6 | 1,145.8 – 1,195.3 | 940 |
| 159 | Grok Code Fast 1 | xai | 1,165.9 | 1,137.5 – 1,194.3 | 780 |
| 160 | Grok 4 Fast (Reasoning) | xai | 1,159.1 | 1,131.6 – 1,186.5 | 742 |
| 161 | Devstral Medium (2507) | mistral | 1,072.0 | 1,041.3 – 1,102.7 | 834 |
数据来源:Arena 官方网页开发。 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 18:06(UTC+8)。