AGI-Eval Community - Mathematical Reasoning: leaderboard
AGI-Eval community leaderboard (agi-eval.cn), the private-set LLM board built from the platform's non-public items; latest monthly snapshot. Unrelated to the AGIEval exam dataset. This board is the mathematical reasoning dimension; upstream's Chinese split is identical, so only this column is kept. Upstream accuracy 0-1, shown as 0-100; higher is better.
Metric: Accuracy (%). Source: agi-eval.cn. Saturation forecast: Estimated already saturated. 141 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 | 89.37 | #20 |
| 2 | Claude Opus 4.8 | 88.21 | #39 |
| 3 | Qwen 3.7 Max | 88.11 | #85 |
| 4 | Kimi K2.6 | 88.04 | #113 |
| 5 | Seed 2.0 Pro | 87.58 | #122 |
| 6 | Claude Opus 4.5 | 87.18 | #87 |
| 7 | Claude Sonnet 5 | 86.98 | #65 |
| 8 | DeepSeek V4 Flash (Max) | 86.96 | #155 (DeepSeek V4 Flash) |
| 9 | Claude Opus 4.6 | 86.76 | #67 |
| 10 | Kimi K2.5 (Thinking) | 86.75 | #152 (Kimi K2.5) |
| 11 | Gemini 3.5 Flash | 86.74 | #66 |
| 12 | Qwen 3 Max (Thinking) | 86.73 | #213 (Qwen 3 Max) |
| 13 | Hy3 | 86.59 | #100 |
| 14 | Gemini 3.1 Pro (Preview) | 86.55 | #64 |
| 15 | DeepSeek V4 Pro (Max) | 86.24 | #112 (DeepSeek V4 Pro) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=agi-eval-community-mathematical-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-05.