AGI-Eval Community - Learning (Chinese): leaderboard
Metric: Accuracy (%). Source: agi-eval.cn. 139 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 91.15 |
| 2 | Gemini 3 Flash (Preview) | 89.43 |
| 3 | Gemini 3 Pro (Preview) | 87.49 |
| 4 | Gemini 3.5 Flash | 87.2 |
| 5 | GLM-5.1 | 87.13 |
| 6 | GLM-5 | 86.81 |
| 7 | GPT-5.5 | 86.42 |
| 8 | Claude Opus 4.6 | 86.3 |
| 9 | Claude Fable 5 | 86.16 |
| 10 | GLM-5.2 | 86.08 |
| 11 | DeepSeek V4 Pro (Max) | 85.9 |
| 12 | Gemini 2.5 Flash | 85.84 |
| 13 | Qwen 3.7 Max | 85.73 |
| 14 | GLM-4.7 | 85.67 |
| 15 | Kimi K2.5 | 85.64 |
Interactive version: theaggregate.ai/benchmark?slug=agi-eval-community-learning-chinese · How It Works · Data refreshed daily, snapshot 2026-09-19.