AGI-Eval Community - General 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 general reasoning dimension; its English split carries about 93% of the weight and its small Chinese split is saturated, 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 | 97.82 | #20 |
| 2 | Gemini 3.1 Pro (Preview) | 97.68 | #64 |
| 3 | Hy3-preview | 97.68 | #166 |
| 4 | Claude Opus 4.6 | 97.66 | #67 |
| 5 | Qwen 3.7 Max | 97.56 | #85 |
| 6 | Gemini 3 Pro (Preview) | 97.51 | #71 |
| 7 | Kimi K3 | 97.5 | #35 |
| 8 | Ring-1T | 97.49 | #326 |
| 9 | Hy3 | 97.48 | #100 |
| 10 | Seed 2.1 Pro | 97.48 | #73 |
| 11 | Claude Opus 4 (Thinking) | 97.47 | #170 (Claude Opus 4) |
| 12 | Kimi K2.6 | 97.45 | #113 |
| 13 | Gemini 3.5 Flash | 97.45 | #66 |
| 14 | ERNIE 5.0 Thinking Preview | 97.41 | #240 |
| 15 | DeepSeek V4 Pro (Max) | 97.38 | #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-general-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-05.