AGI-Eval Community - Learning: 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 learning dimension, combining Chinese and English items. 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 | Gemini 3.1 Pro (Preview) | 94.78 | #64 |
| 2 | Gemini 3 Flash (Preview) | 93.89 | #90 |
| 3 | Gemini 3.5 Flash | 92.41 | #66 |
| 4 | Gemini 3 Pro (Preview) | 92.27 | #71 |
| 5 | GLM-5.1 | 92.26 | #109 |
| 6 | Qwen 3.7 Max | 92.03 | #85 |
| 7 | GPT-5.5 | 91.89 | #33 |
| 8 | GLM-5.2 | 91.86 | #80 |
| 9 | DeepSeek V4 Pro (Max) | 91.55 | #112 (DeepSeek V4 Pro) |
| 10 | GLM-5 | 91.4 | #149 |
| 11 | Claude Opus 4.6 | 91.22 | #67 |
| 12 | Claude Opus 4.8 | 91.13 | #39 |
| 13 | GLM-4.7 | 91.02 | #202 |
| 14 | Qwen 3.5 397B A17B (Thinking) | 90.99 | #154 (Qwen 3.5 397B A17B) |
| 15 | Kimi K3 | 90.9 | #35 |
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-learning · How It Works · Data refreshed daily, snapshot 2026-11-05.