这个评测关注什么
覆盖真实职业任务和高经济价值知识工作。
阅读时注意
评测结果受任务集合、评分方法和模型版本影响。本站保留原始口径,不用它生成新的综合排名。
适用场景
用于回答一个明确的问题:模型在这一类任务上的相对表现如何,而不是“哪个模型绝对最好”。
经济价值任务
GDPval-AA v2 公开评测榜单:覆盖真实职业任务和高经济价值知识工作。本站保留来源口径,不合成新的综合评分。
| 排名 | 机构 | 模型 | Elo | 95% CI | 发布日期 |
|---|---|---|---|---|---|
| 1 | Anthropic | Claude Fable 5.1 (Adaptive Reasoning, Xhigh Effort, Default Fallback) | 1,745 裁判协议 v4.3.1 | -19 / +19 | Sep 2026 |
| 2 | Anthropic | Claude Opus 5 (Adaptive Reasoning, Max Effort) | 1,735 裁判协议 v4.3.1 | -16 / +16 | Jul 2026 |
| 3 | Anthropic | Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback) | 1,724 裁判协议 v4.3.1 | -16 / +16 | Sep 2026 |
| 4 | Anthropic | Claude Opus 5 (Adaptive Reasoning, Xhigh Effort) | 1,708 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 |
| 5 | Meta | Muse Spark 1.3 (max) | 1,703 裁判协议 v4.3.1 | -19 / +19 | Sep 2026 |
| 6 | Alibaba | Qwen3.8 Max (0902) | 1,664 裁判协议 v4.3.1 | -19 / +19 | Sep 2026 |
| 7 | SpaceXAI | Grok 4.6 (xhigh) | 1,663 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 |
| 8 | Meta | Muse Spark 1.3 (xhigh) | 1,662 裁判协议 v4.3.1 | -19 / +19 | Sep 2026 |
| 9 | Z AI | GLM 5.3 Flash | 1,655 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 |
| 10 | Anthropic | Claude Fable 5.1 (Adaptive Reasoning, High Effort, Default Fallback) | 1,650 裁判协议 v4.3.1 | -18 / +18 | Sep 2026 |
| 11 | Alibaba | Qwen3.8-Flash-Next | 1,647 裁判协议 v4.3.1 | -17 / +17 | Aug 2026 |
| 12 | SpaceXAI | Grok 4.6 (high) | 1,643 裁判协议 v4.3.1 | -14 / +14 | Aug 2026 |
| 13 | SpaceXAI | Grok 4.6 (medium) | 1,643 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 |
| 14 | Z AI | GLM-5.3 (max) | 1,634 裁判协议 v4.3.1 | -14 / +14 | Aug 2026 |
| 15 | DeepSeek | DeepSeek V4.1 Flash (Reasoning, Max Effort) | 1,632 裁判协议 v4.3.1 | -21 / +21 | Sep 2026 |
| 16 | Anthropic | Claude Fable 5 | 1,631 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 17 | Alibaba | Qwen3.8 Max | 1,630 裁判协议 v4.3.1 | -14 / +14 | Aug 2026 |
| 18 | Anthropic | Claude Opus 5 (Adaptive Reasoning, High Effort) | 1,629 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 |
| 19 | Alibaba | Qwen3.8 2.4T A95B | 1,628 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 |
| 20 | OpenAI | GPT-5.6 Sol (max) | 1,587 裁判协议 v4.3.1 | -13 / +13 | Jul 2026 |
| 21 | OpenAI | GPT-5.6 Sol (xhigh) | 1,585 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 22 | OpenAI | GPT-6 Astra (max) | 1,580 裁判协议 v4.3.1 | -20 / +20 | Sep 2026 |
| 23 | Anthropic | Claude Fable 5.1 (Adaptive Reasoning, Medium Effort, Default Fallback) | 1,579 裁判协议 v4.3.1 | -18 / +18 | Sep 2026 |
| 24 | DeepSeek | DeepSeek V4 Flash Vision (Reasoning, Max Effort) | 1,577 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 |
| 25 | Sapiens AI | Agnes 3.0 Flash | 1,573 裁判协议 v4.3.1 | -22 / +22 | Sep 2026 |
| 26 | OpenAI | GPT-6 Astra (xhigh) | 1,555 裁判协议 v4.3.1 | -20 / +20 | Sep 2026 |
| 27 | Kimi | Kimi K3 (max) | 1,548 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 28 | OpenAI | GPT-6 Astra (high) | 1,530 裁判协议 v4.3.1 | -20 / +20 | Sep 2026 |
| 29 | Anthropic | Claude Opus 5 (Adaptive Reasoning, Medium Effort) | 1,525 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 |
| 30 | OpenAI | GPT-5.6 Sol (high) | 1,524 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 31 | Meta | Muse Spark 1.2 (xhigh) | 1,523 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 |
| 32 | Anthropic | Claude Fable 5.1 (Adaptive Reasoning, Low Effort, Default Fallback) | 1,504 裁判协议 v4.3.1 | -18 / +18 | Sep 2026 |
| 33 | Anthropic | Claude Sonnet 5 (Adaptive Reasoning, Max Effort) | 1,501 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 34 | OpenAI | GPT-6 Astra (medium) | 1,501 裁判协议 v4.3.1 | -20 / +20 | Sep 2026 |
| 35 | DeepSeek | DeepSeek V4 Pro 0813 (Reasoning, Max Effort) | 1,493 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 |
| 36 | Anthropic | Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | 1,489 裁判协议 v4.3.1 | -13 / +13 | May 2026 |
| 37 | OpenAI | GPT-5.6 Terra (xhigh) | 1,479 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 38 | OpenAI | GPT-5.6 Terra (max) | 1,477 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 39 | Gemini 3.8 Flash (high) | 1,464 裁判协议 v4.3.1 | -19 / +19 | Sep 2026 | |
| 40 | Alibaba | Qwen3.8 27B (xhigh) | 1,463 裁判协议 v4.3.1 | -16 / +16 | Aug 2026 |
| 41 | SpaceXAI | Grok 4.6 (low) | 1,461 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 |
| 42 | OpenAI | GPT-5.6 Luna (max) | 1,456 裁判协议 v4.3.1 | -13 / +13 | Jul 2026 |
| 43 | Gemini 3.8 Flash (medium) | 1,455 裁判协议 v4.3.1 | -19 / +19 | Sep 2026 | |
| 44 | DeepSeek | DeepSeek V4 Flash 0731 (Reasoning, Max Effort) | 1,442 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 45 | Gemini 3.7 Flash (high) | 1,433 裁判协议 v4.3.1 | -15 / +15 | Aug 2026 | |
| 46 | Alibaba | Qwen3.8 27B (medium) | 1,430 裁判协议 v4.3.1 | -16 / +16 | Aug 2026 |
| 47 | SpaceXAI | Grok 4.5 (high) | 1,430 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 |
| 48 | OpenAI | GPT-5.6 Luna (xhigh) | 1,428 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 49 | OpenAI | GPT-5.6 Sol (medium) | 1,427 裁判协议 v4.3.1 | -13 / +13 | Jul 2026 |
| 50 | OpenAI | GPT-6 Astra (low) | 1,419 裁判协议 v4.3.1 | -20 / +20 | Sep 2026 |
| 51 | OpenAI | GPT-5.6 Terra (high) | 1,415 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 52 | Gemini 3.7 Flash (medium) | 1,407 裁判协议 v4.3.1 | -16 / +16 | Aug 2026 | |
| 53 | Z AI | GLM-5.2 (max) | 1,406 裁判协议 v4.3.1 | -13 / +13 | Jun 2026 |
| 54 | Alibaba | Qwen3.8 27B (low) | 1,403 裁判协议 v4.3.1 | -16 / +16 | Aug 2026 |
| 55 | Anthropic | Claude Sonnet 5 (Adaptive Reasoning, Xhigh Effort) | 1,401 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 56 | Institute of Foundation Models | K2 Horizon 375B A23B | 1,400 裁判协议 v4.3.1 | -19 / +19 | Sep 2026 |
| 57 | OpenAI | GPT-5.5 (xhigh) | 1,396 裁判协议 v4.3.1 | -13 / +13 | Apr 2026 |
| 58 | Anthropic | Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | 1,396 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 59 | Sapiens AI | Agnes 2.5 Pro Beta | 1,376 裁判协议 v4.3.1 | -17 / +17 | Aug 2026 |
| 60 | OpenAI | GPT-5.6 Luna (high) | 1,375 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 61 | Anthropic | Claude Opus 5 (Adaptive Reasoning, Low Effort) | 1,372 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 |
| 62 | OpenAI | GPT-5.5 (high) | 1,372 裁判协议 v4.3.1 | -13 / +13 | Apr 2026 |
| 63 | Gemini 3.8 Flash (low) | 1,359 裁判协议 v4.3.1 | -16 / +16 | Sep 2026 | |
| 64 | OpenAI | GPT-5.6 Sol (low) | 1,355 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 65 | Gemini 3.7 Flash (low) | 1,354 裁判协议 v4.3.1 | -16 / +16 | Aug 2026 | |
| 66 | Gemini 3.6 Flash (high) | 1,331 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 | |
| 67 | OpenAI | GPT-5.6 Terra (medium) | 1,318 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 68 | Anthropic | Claude Sonnet 5 (Adaptive Reasoning, High Effort) | 1,308 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 69 | OpenAI | GPT-5.4 (xhigh) | 1,307 裁判协议 v4.3.1 | -14 / +14 | Mar 2026 |
| 70 | Z AI | GLM-5.2 (Non-reasoning) | 1,306 裁判协议 v4.3.1 | -16 / +16 | Jun 2026 |
| 71 | MiniMax | MiniMax-M3 | 1,304 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 72 | Anthropic | Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | 1,295 裁判协议 v4.3.1 | -14 / +14 | Feb 2026 |
| 73 | Meta | Muse Spark 1.1 (xhigh) | 1,294 裁判协议 v4.3.1 | -16 / +16 | Jul 2026 |
| 74 | OpenAI | GPT-5.5 (medium) | 1,288 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 75 | Anthropic | Claude Sonnet 5 (Non-reasoning, High Effort) | 1,283 裁判协议 v4.3.1 | -15 / +15 | Jun 2026 |
| 76 | OpenAI | GPT-5.6 Sol (Non-reasoning) | 1,273 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 77 | Apodex | Apodex 1.1 | 1,273 裁判协议 v4.3.1 | -17 / +17 | Aug 2026 |
| 78 | Gemini 3.5 Flash (high) | 1,259 裁判协议 v4.3.1 | -14 / +14 | May 2026 | |
| 79 | Institute of Foundation Models | K2 Horizon MoVA 36B A4B | 1,243 裁判协议 v4.3.1 | -21 / +21 | Sep 2026 |
| 80 | Multiverse Computing | Quasar 438B (max, based on GLM-5.2) | 1,239 裁判协议 v4.3.1 | -19 / +19 | Aug 2026 |
| 81 | InclusionAI | Ling-3.0-flash-VL | 1,225 裁判协议 v4.3.1 | -21 / +21 | Sep 2026 |
| 82 | DeepSeek | DeepSeek V4 Pro 0424 (Reasoning, Max Effort) | 1,223 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 83 | Anthropic | Claude Sonnet 5 (Adaptive Reasoning, Medium Effort) | 1,223 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 84 | DeepSeek | DeepSeek V4 Pro 0424 (Reasoning, High Effort) | 1,219 裁判协议 v4.3.1 | -17 / +17 | Apr 2026 |
| 85 | Motif Technologies | Motif 3 | 1,199 裁判协议 v4.3.1 | -17 / +17 | Aug 2026 |
| 86 | China Mobile | JT-4.1 Flash 236B A21B | 1,193 裁判协议 v4.3.1 | -16 / +16 | Jul 2026 |
| 87 | Thinking Machines | Inkling Small | 1,191 裁判协议 v4.3.1 | -17 / +17 | Jul 2026 |
| 88 | Alibaba | Qwen3.7 Max | 1,190 裁判协议 v4.3.1 | -14 / +14 | May 2026 |
| 89 | AI9Stars | G9v3-39A5B | 1,189 裁判协议 v4.3.1 | -18 / +18 | Aug 2026 |
| 90 | Xiaomi | MiMo-V2.5-Pro | 1,186 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 91 | Kimi | Kimi K3 (low) | 1,185 裁判协议 v4.3.1 | -17 / +17 | Jul 2026 |
| 92 | Motif Technologies | Motif 3 (Beta) | 1,183 裁判协议 v4.3.1 | -17 / +17 | Jul 2026 |
| 93 | Z AI | GLM-5.1 (Reasoning) | 1,180 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 94 | OpenAI | GPT-5.6 Terra (low) | 1,178 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 |
| 95 | Nex AGI | Nex-N2-Pro (based on Qwen3.5-397B-A17B) | 1,175 裁判协议 v4.3.1 | -15 / +15 | Jun 2026 |
| 96 | Upstage | Solar Pro 4 | 1,173 裁判协议 v4.3.1 | -16 / +16 | Aug 2026 |
| 97 | InclusionAI | Ling-3.0-flash-Fin | 1,171 裁判协议 v4.3.1 | -21 / +21 | Sep 2026 |
| 98 | Thinking Machines | Inkling (xhigh) | 1,165 裁判协议 v4.3.1 | -17 / +17 | Jul 2026 |
| 99 | OpenAI | GPT-5.6 Luna (medium) | 1,164 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 100 | OpenAI | GPT-5.6 Terra (Non-reasoning) | 1,146 裁判协议 v4.3.1 | -14 / +14 | Jul 2026 |
| 101 | Anthropic | Claude Sonnet 5 (Adaptive Reasoning, Low Effort) | 1,145 裁判协议 v4.3.1 | -15 / +15 | Jun 2026 |
| 102 | SpaceXAI | Grok Build 0.1 0616 | 1,138 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 103 | Tencent | Hy3 | 1,136 裁判协议 v4.3.1 | -17 / +17 | Jul 2026 |
| 104 | DeepSeek | DeepSeek V4 Flash 0420 (Reasoning, Max Effort) | 1,116 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 105 | Kimi | Kimi K2.7 Code | 1,114 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 106 | Alibaba | Qwen3.8 27B (Non-reasoning) | 1,113 裁判协议 v4.3.1 | -17 / +17 | Aug 2026 |
| 107 | OpenAI | GPT-5.5 (low) | 1,113 裁判协议 v4.3.1 | -15 / +15 | Apr 2026 |
| 108 | Kimi | Kimi K2.6 | 1,096 裁判协议 v4.3.1 | -13 / +13 | Apr 2026 |
| 109 | Z AI | GLM-4.7 (Reasoning) | 1,096 裁判协议 v4.3.1 | -16 / +16 | Dec 2025 |
| 110 | OpenAI | GPT-5.4 mini (xhigh) | 1,095 裁判协议 v4.3.1 | -14 / +14 | Mar 2026 |
| 111 | Institute of Foundation Models | K2 Horizon 7B | 1,091 裁判协议 v4.3.1 | -21 / +21 | Sep 2026 |
| 112 | NVIDIA | Nemotron 3 Ultra 550B A55B (Reasoning) | 1,091 裁判协议 v4.3.1 | -14 / +14 | Jun 2026 |
| 113 | Sapiens AI | Agnes 2.5 Pro Alpha | 1,088 裁判协议 v4.3.1 | -17 / +17 | Jul 2026 |
| 114 | MiniMax | MiniMax-M2.7 | 1,087 裁判协议 v4.3.1 | -14 / +14 | Mar 2026 |
| 115 | DeepSeek | DeepSeek V4 Flash 0420 (Reasoning, High Effort) | 1,083 裁判协议 v4.3.1 | -17 / +17 | Apr 2026 |
| 116 | Xiaomi | MiMo-V2.5 | 1,079 裁判协议 v4.3.1 | -17 / +17 | Apr 2026 |
| 117 | Meta | Muse Spark | 1,076 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 118 | OpenAI | GPT-5.6 Luna (low) | 1,075 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 |
| 119 | Alibaba | Qwen3.6 27B (Reasoning) | 1,069 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 120 | Alibaba | Qwen3.6 Plus | 1,066 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 121 | Gemini 3.5 Flash-Lite | 1,063 裁判协议 v4.3.1 | -17 / +17 | Jul 2026 | |
| 122 | OpenAI | GPT-5.5 (Non-reasoning) | 1,054 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 123 | Upstage | Solar Open2 250B | 1,053 裁判协议 v4.3.1 | -18 / +18 | Aug 2026 |
| 124 | SK Telecom | A.X-K2 | 1,044 裁判协议 v4.3.1 | -18 / +18 | Aug 2026 |
| 125 | Alibaba | Qwen3.6 27B (Non-reasoning) | 1,043 裁判协议 v4.3.1 | -16 / +16 | Apr 2026 |
| 126 | OpenAI | GPT-5.4 nano (xhigh) | 1,035 裁判协议 v4.3.1 | -14 / +14 | Mar 2026 |
| 127 | SpaceXAI | Grok 4.3 (Non-reasoning) | 1,031 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 128 | SpaceXAI | Grok 4.3 (high) | 1,018 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 129 | InclusionAI | Ling 3.0 Flash | 1,017 裁判协议 v4.3.1 | -17 / +17 | Aug 2026 |
| 130 | OpenAI | GPT-5 (high) | 1,015 裁判协议 v4.3.1 | -16 / +16 | Aug 2025 |
| 131 | Alibaba | Qwen3.6 35B A3B (Reasoning) | 991 裁判协议 v4.3.1 | -14 / +14 | Apr 2026 |
| 132 | Anthropic | Claude 4.5 Sonnet (Reasoning) | 989 裁判协议 v4.3.1 | -16 / +16 | Sep 2025 |
| 133 | OpenAI | GPT-5.6 Luna (Non-reasoning) | 987 裁判协议 v4.3.1 | -15 / +15 | Jul 2026 |
| 134 | LongCat | LongCat 2.0 | 965 裁判协议 v4.3.1 | -16 / +16 | Jun 2026 |
| 135 | Institute of Foundation Models | K2 Horizon 3.7B | 962 裁判协议 v4.3.1 | -21 / +21 | Sep 2026 |
| 136 | StepFun | Step 3.7 Flash | 954 裁判协议 v4.3.1 | -14 / +14 | May 2026 |
| 137 | Alibaba | Qwen3.6 35B A3B (Non-reasoning) | 953 裁判协议 v4.3.1 | -18 / +18 | Apr 2026 |
| 138 | Kimi | Kimi K2.5 (Reasoning) | 936 裁判协议 v4.3.1 | -17 / +17 | Jan 2026 |
| 139 | OpenAI | GPT-5.1 (high) | 930 裁判协议 v4.3.1 | -17 / +17 | Nov 2025 |
| 140 | Alibaba | Qwen3.5 122B A10B (Reasoning) | 925 裁判协议 v4.3.1 | -15 / +15 | Feb 2026 |
| 141 | LG AI Research | K-EXAONE 2.0 0803 | 918 裁判协议 v4.3.1 | -19 / +19 | Aug 2026 |
| 142 | Alibaba | Qwen3.5 397B A17B (Reasoning) | 905 裁判协议 v4.3.1 | -15 / +15 | Feb 2026 |
| 143 | Gemini 3.1 Pro Preview | 904 裁判协议 v4.3.1 | -15 / +15 | Feb 2026 | |
| 144 | Meta | Muse Glimmer (high) | 893 裁判协议 v4.3.1 | -19 / +19 | Aug 2026 |
| 145 | Alibaba | Qwen3.7 Plus | 886 裁判协议 v4.3.1 | -15 / +15 | Jun 2026 |
| 146 | OpenAI | GPT-5 mini (high) | 880 裁判协议 v4.3.1 | -17 / +17 | Aug 2025 |
| 147 | Mistral | Mistral Medium 3.5 | 875 裁判协议 v4.3.1 | -15 / +15 | Apr 2026 |
| 148 | Z AI | GLM-4.6 (Reasoning) | 874 裁判协议 v4.3.1 | -17 / +17 | Sep 2025 |
| 149 | InclusionAI | Ring-2.6-1T | 864 裁判协议 v4.3.1 | -15 / +15 | May 2026 |
| 150 | Anthropic | Claude 4.5 Haiku (Reasoning) | 854 裁判协议 v4.3.1 | -15 / +15 | Oct 2025 |
| 151 | KwaiKAT | KAT-Coder-Pro V1 | 845 裁判协议 v4.3.1 | -16 / +16 | Nov 2025 |
| 152 | KwaiKAT | KAT Coder Pro V2 | 843 裁判协议 v4.3.1 | -18 / +18 | Mar 2026 |
| 153 | Alibaba | Qwen3.5 122B A10B (Non-reasoning) | 833 裁判协议 v4.3.1 | -18 / +18 | Feb 2026 |
| 154 | OpenBMB | MiniCPM5-2B | 828 裁判协议 v4.3.1 | -22 / +22 | Sep 2026 |
| 155 | DeepSeek | DeepSeek V3.1 Terminus (Reasoning) | 826 裁判协议 v4.3.1 | -18 / +18 | Sep 2025 |
| 156 | Anthropic | Claude 4 Sonnet (Reasoning) | 819 裁判协议 v4.3.1 | -17 / +17 | May 2025 |
| 157 | DeepSeek | DeepSeek V3.2 (Reasoning) | 810 裁判协议 v4.3.1 | -19 / +19 | Dec 2025 |
| 158 | AI9Stars | G9v3-3B | 807 裁判协议 v4.3.1 | -26 / +26 | Jul 2026 |
| 159 | Xiaomi | MiMo-V2-Flash (Non-reasoning) | 788 裁判协议 v4.3.1 | -19 / +19 | Dec 2025 |
| 160 | NVIDIA | Nemotron 3.5 Lightning | 767 裁判协议 v4.3.1 | -19 / +19 | Aug 2026 |
| 161 | Gemma 4 31B (Reasoning) | 755 裁判协议 v4.3.1 | -16 / +16 | Apr 2026 | |
| 162 | OpenAI | gpt-oss-120b (high) | 745 裁判协议 v4.3.1 | -15 / +15 | Aug 2025 |
| 163 | Alibaba | Qwen3.5 35B A3B (Non-reasoning) | 740 裁判协议 v4.3.1 | -20 / +20 | Feb 2026 |
| 164 | OpenAI | GPT-5.4 mini (Non-Reasoning) | 734 裁判协议 v4.3.1 | -16 / +16 | Mar 2026 |
| 165 | IBM | Granite 4.2 30B | 721 裁判协议 v4.3.1 | -20 / +20 | Aug 2026 |
| 166 | InclusionAI | Ling 3.0 Tiny | 716 裁判协议 v4.3.1 | -20 / +20 | Aug 2026 |
| 167 | Gemma 4 26B A4B (Reasoning) | 693 裁判协议 v4.3.1 | -15 / +15 | Apr 2026 | |
| 168 | Gemma 4 31B (Non-reasoning) | 691 裁判协议 v4.3.1 | -18 / +18 | Apr 2026 | |
| 169 | Mistral | Devstral 2 | 689 裁判协议 v4.3.1 | -17 / +17 | Dec 2025 |
| 170 | OpenAI | GPT-5.4 nano (Non-Reasoning) | 681 裁判协议 v4.3.1 | -21 / +21 | Mar 2026 |
| 171 | Mistral | Devstral Small 2 | 680 裁判协议 v4.3.1 | -16 / +16 | Dec 2025 |
| 172 | Alibaba | Qwen3 Coder Next | 664 裁判协议 v4.3.1 | -18 / +18 | Feb 2026 |
| 173 | OpenAI | GPT-5.5 Instant (June 2026) | 662 裁判协议 v4.3.1 | -17 / +17 | Jun 2026 |
| 174 | Cohere | Command A+ | 658 裁判协议 v4.3.1 | -18 / +18 | May 2026 |
| 175 | Inception | Mercury 2 | 646 裁判协议 v4.3.1 | -19 / +19 | Feb 2026 |
| 176 | IBM | Granite 4.2 8B | 645 裁判协议 v4.3.1 | -21 / +21 | Aug 2026 |
| 177 | NVIDIA | Nemotron 3 Super 120B A12B (Reasoning) | 644 裁判协议 v4.3.1 | -16 / +16 | Mar 2026 |
| 178 | Amazon | Nova 2.0 Pro Preview (medium) | 630 裁判协议 v4.3.1 | -16 / +16 | Nov 2025 |
| 179 | LG AI Research | EXAONE 4.5 33B | 628 裁判协议 v4.3.1 | -20 / +20 | Apr 2026 |
| 180 | Gemini 2.5 Pro | 616 裁判协议 v4.3.1 | -17 / +17 | Jun 2025 | |
| 181 | Multiverse Computing | HyperNova 60B 2605 (high, based on gpt-oss-120b) | 604 裁判协议 v4.3.1 | -19 / +19 | May 2026 |
| 182 | Amazon | Nova 2.0 Pro Preview (low) | 601 裁判协议 v4.3.1 | -16 / +16 | Nov 2025 |
| 183 | Gemini 3.1 Flash-Lite | 596 裁判协议 v4.3.1 | -16 / +16 | Mar 2026 | |
| 184 | Gemma 4 12B (Reasoning) | 591 裁判协议 v4.3.1 | -20 / +20 | Jun 2026 | |
| 185 | Alibaba | Qwen3.5 9B (Reasoning) | 588 裁判协议 v4.3.1 | -18 / +18 | Mar 2026 |
| 186 | Mistral | Mistral Large 3 | 588 裁判协议 v4.3.1 | -16 / +16 | Dec 2025 |
| 187 | Mistral | Mistral Small 3.1 | 550 裁判协议 v4.3.1 | -18 / +18 | Mar 2025 |
| 188 | LG AI Research | K-EXAONE (Reasoning) | 540 裁判协议 v4.3.1 | -20 / +20 | Dec 2025 |
| 189 | Mistral | Mistral Small 4 (Reasoning) | 537 裁判协议 v4.3.1 | -19 / +19 | Mar 2026 |
| 190 | Mistral | Mistral Medium 3.1 | 536 裁判协议 v4.3.1 | -17 / +17 | Aug 2025 |
| 191 | Amazon | Nova 2.0 Lite (high) | 536 裁判协议 v4.3.1 | -17 / +17 | Oct 2025 |
| 192 | OpenAI | gpt-oss-20b (high) | 514 裁判协议 v4.3.1 | -16 / +16 | Aug 2025 |
| 193 | Arcee AI | Trinity Large Thinking | 512 裁判协议 v4.3.1 | -20 / +20 | Apr 2026 |
| 194 | Amazon | Nova 2.0 Pro Preview (Non-reasoning) | 509 裁判协议 v4.3.1 | -16 / +16 | Nov 2025 |
| 195 | DiffusionGemma 26B A4B | 501 裁判协议 v4.3.1 | -17 / +17 | Jun 2026 | |
| 196 | Alibaba | Qwen3 235B A22B 2507 (Reasoning) | 495 裁判协议 v4.3.1 | -19 / +19 | Jul 2025 |
| 197 | InclusionAI | Ling 2.6 Flash | 493 裁判协议 v4.3.1 | -19 / +19 | Apr 2026 |
| 198 | Celeris | Celeris-1 | 483 裁判协议 v4.3.1 | -21 / +21 | Jul 2026 |
| 199 | DeepSeek | DeepSeek R1 (Jan '25) | 479 裁判协议 v4.3.1 | -18 / +18 | Jan 2025 |
| 200 | Cohere | North Mini Code | 472 裁判协议 v4.3.1 | -18 / +18 | Jun 2026 |
| 201 | OpenAI | GPT-4.1 mini | 453 裁判协议 v4.3.1 | -18 / +18 | Apr 2025 |
| 202 | NVIDIA | Nemotron Cascade 2 30B A3B | 447 裁判协议 v4.3.1 | -21 / +21 | Mar 2026 |
| 203 | Upstage | Solar Pro 3 | 447 裁判协议 v4.3.1 | -16 / +16 | Apr 2026 |
| 204 | NVIDIA | NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) | 438 裁判协议 v4.3.1 | -17 / +17 | Dec 2025 |
| 205 | Mistral | Ministral 3 14B | 432 裁判协议 v4.3.1 | -17 / +17 | Dec 2025 |
| 206 | OpenAI | o3-mini (high) | 421 裁判协议 v4.3.1 | -19 / +19 | Jan 2025 |
| 207 | NVIDIA | Nemotron 3 Nano Omni 30B A3B Reasoning | 416 裁判协议 v4.3.1 | -21 / +21 | Apr 2026 |
| 208 | Anthropic | Claude 3.5 Haiku | 408 裁判协议 v4.3.1 | -18 / +18 | Oct 2024 |
| 209 | Mistral | Ministral 3 8B | 405 裁判协议 v4.3.1 | -17 / +17 | Dec 2025 |
| 210 | IBM | Granite 4.2 3B | 390 裁判协议 v4.3.1 | -21 / +21 | Aug 2026 |
| 211 | IBM | Granite 4.1 30B | 379 裁判协议 v4.3.1 | -16 / +16 | Apr 2026 |
| 212 | OpenAI | gpt-oss-120b (low) | 364 裁判协议 v4.3.1 | -21 / +21 | Aug 2025 |
| 213 | Mistral | Magistral Medium 1.2 | 362 裁判协议 v4.3.1 | -18 / +18 | Sep 2025 |
| 214 | Institute of Foundation Models | K2 Think V2 | 328 裁判协议 v4.3.1 | -18 / +18 | Dec 2025 |
| 215 | Alibaba | Qwen3 Next 80B A3B (Reasoning) | 325 裁判协议 v4.3.1 | -20 / +20 | Sep 2025 |
| 216 | DeepSeek | DeepSeek V3 0324 | 275 裁判协议 v4.3.1 | -19 / +19 | Mar 2025 |
| 217 | Alibaba | Qwen3 30B A3B 2507 (Reasoning) | 274 裁判协议 v4.3.1 | -20 / +20 | Jul 2025 |
| 218 | Alibaba | Qwen3 32B (Reasoning) | 240 裁判协议 v4.3.1 | -18 / +18 | Apr 2025 |
| 219 | Mistral | Ministral 3 3B | 234 裁判协议 v4.3.1 | -18 / +18 | Dec 2025 |
| 220 | Mistral | Magistral Small 1.2 | 213 裁判协议 v4.3.1 | -19 / +19 | Sep 2025 |
| 221 | Liquid AI | LFM2.5-2.6B | 202 裁判协议 v4.3.1 | -22 / +22 | Aug 2026 |
| 222 | OpenAI | GPT-4 | 189 裁判协议 v4.3.1 | -18 / +18 | Mar 2023 |
| 223 | OpenAI | GPT-4o mini | 188 裁判协议 v4.3.1 | -19 / +19 | Jul 2024 |
| 224 | Alibaba | Qwen3 14B (Reasoning) | 188 裁判协议 v4.3.1 | -19 / +19 | Apr 2025 |
| 225 | DeepSeek | DeepSeek V3 (Dec '24) | 183 裁判协议 v4.3.1 | -19 / +19 | Dec 2024 |
| 226 | Gemma 4 E4B (Reasoning) | 177 裁判协议 v4.3.1 | -22 / +22 | Apr 2026 | |
| 227 | Alibaba | Qwen3.5 2B (Reasoning) | 166 裁判协议 v4.3.1 | -21 / +21 | Mar 2026 |
| 228 | Alibaba | Qwen3 8B (Reasoning) | 165 裁判协议 v4.3.1 | -19 / +19 | Apr 2025 |
| 229 | NVIDIA | NVIDIA Nemotron 3 Nano 4B | 159 裁判协议 v4.3.1 | -21 / +21 | Mar 2026 |
| 230 | IBM | Granite 4.1 3B | 74 裁判协议 v4.3.1 | -20 / +20 | Apr 2026 |
| 231 | Meta | Llama 4 Scout | 60 裁判协议 v4.3.1 | -17 / +17 | Apr 2025 |
| 232 | Meta | Llama 3.3 Instruct 70B | 50 裁判协议 v4.3.1 | -19 / +19 | Dec 2024 |
| 233 | Gemma 4 E2B (Reasoning) | 36 裁判协议 v4.3.1 | -21 / +21 | Apr 2026 | |
| 234 | Mistral | Mistral Small 3.2 | 29 裁判协议 v4.3.1 | -19 / +19 | Jun 2025 |
| 235 | OpenAI | GPT-4.1 nano | 13 裁判协议 v4.3.1 | -19 / +19 | Apr 2025 |
| 236 | Meta | Llama 4 Maverick | -45 裁判协议 v4.3.1 | -17 / +17 | Apr 2025 |
| 237 | NVIDIA | NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning) | -122 裁判协议 v4.3.1 | -17 / +17 | Dec 2025 |
| 238 | Alibaba | Qwen3.5 2B (Non-reasoning) | -130 裁判协议 v4.3.1 | -18 / +18 | Mar 2026 |
| 239 | Alibaba | Qwen3.5 0.8B (Non-reasoning) | -133 裁判协议 v4.3.1 | -18 / +18 | Mar 2026 |
| 240 | OpenBMB | MiniCPM-V 4.6 1.3B | -134 裁判协议 v4.3.1 | -17 / +17 | May 2026 |
| 241 | Meta | Llama 3.1 Instruct 8B | -151 裁判协议 v4.3.1 | -17 / +17 | Jul 2024 |
| 242 | Alibaba | Qwen3.5 0.8B (Reasoning) | -152 裁判协议 v4.3.1 | -17 / +17 | Mar 2026 |
| 243 | Nanbeige | Nanbeige4.1-3B | -166 裁判协议 v4.3.1 | -17 / +17 | Feb 2026 |
| 244 | Gemma 3 12B Instruct | -170 裁判协议 v4.3.1 | -17 / +17 | Mar 2025 | |
| 245 | Gemma 3 27B Instruct | -171 裁判协议 v4.3.1 | -16 / +16 | Mar 2025 | |
| 246 | Microsoft | Phi-4 Mini Instruct | -171 裁判协议 v4.3.1 | -17 / +17 | Feb 2024 |
| 247 | Institute of Foundation Models | K2 Horizon 0.9B | -174 裁判协议 v4.3.1 | -17 / +17 | Sep 2026 |
| — | OpenAI | GPT-6 Astra (Non-reasoning) 本轮来源未列出,保留上次观测 | 1,543 历史参考 2026-09-12 | -20 / +20 历史参考 2026-09-12 | Sep 2026 |
AA-Briefcase 与 GDPval-AA v2 的当前成对比较使用裁判协议 v4.3.1,原评测名称与任务版本不变,Intelligence Index 仍为 v4.3。未区分裁判协议的旧 Elo 保留原始观测,仅作历史参考,不与当前协议混排。
指数中的 Elo 贡献在模型加入时冻结;实时榜单分数不能用于重算指数。来源未给出逐模型重跑日期,不能把裁判更新或跨协议分数差解释成新的模型运行或能力变化。
查看裁判、核验时间与完整方法边界 →查看 AA 原始方法 ↗覆盖真实职业任务和高经济价值知识工作。
评测结果受任务集合、评分方法和模型版本影响。本站保留原始口径,不用它生成新的综合排名。
用于回答一个明确的问题:模型在这一类任务上的相对表现如何,而不是“哪个模型绝对最好”。