AGI-Eval Community - Subject Reasoning (English): 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 English-language item split (_英) of the subject reasoning column; the parent column combines both splits. Upstream accuracy 0-1, shown as 0-100; higher is better.
Metric: Accuracy (%). Source: agi-eval.cn. Saturation forecast: Estimated already saturated. 139 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 88.79 | #39 |
| 2 | GPT-5.5 | 87.52 | #33 |
| 3 | Claude Fable 5 | 87.32 | #20 |
| 4 | Claude Opus 4.6 | 87.21 | #67 |
| 5 | Gemini 3.5 Flash | 87.16 | #66 |
| 6 | Seed 2.0 Pro | 87.11 | #122 |
| 7 | Qwen 3.7 Max | 87.05 | #85 |
| 8 | Hy3 | 87 | #100 |
| 9 | Kimi K2.5 (Thinking) | 86.9 | #152 (Kimi K2.5) |
| 10 | Kimi K3 | 86.89 | #35 |
| 11 | Kimi K2.6 | 86.79 | #113 |
| 12 | Gemini 3 Pro (Preview) | 86.63 | #71 |
| 13 | Gemini 3.1 Pro (Preview) | 86.53 | #64 |
| 14 | Seed 2.1 Pro | 86.37 | #73 |
| 15 | GPT-5.1 | 86.16 | #145 |
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-subject-reasoning-english · How It Works · Data refreshed daily, snapshot 2026-11-05.