AGI-Eval Community - Subject Reasoning (Chinese): leaderboard
Metric: Accuracy (%). Source: agi-eval.cn. 139 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Seed 2.1 Pro | 94.54 |
| 2 | GPT-5.5 | 94.31 |
| 3 | Seed 2.0 Pro | 94.07 |
| 4 | Claude Fable 5 | 93.6 |
| 5 | Gemini 3.1 Pro (Preview) | 93.36 |
| 6 | Kimi K3 | 93.24 |
| 7 | Gemini 3 Pro (Preview) | 92.88 |
| 8 | Gemini 3.5 Flash | 92.88 |
| 9 | Qwen 3.7 Max | 92.88 |
| 10 | DeepSeek V4 Flash (Max) | 92.88 |
| 11 | Claude Opus 4.8 | 92.76 |
| 12 | Kimi K2.6 | 92.53 |
| 13 | DeepSeek V4 Pro (Max) | 92.53 |
| 14 | Claude Opus 4.6 | 92.41 |
| 15 | Seed 1.8 | 92.29 |
Interactive version: theaggregate.ai/benchmark?slug=agi-eval-community-subject-reasoning-chinese · How It Works · Data refreshed daily, snapshot 2026-09-19.