SuperCLUE: leaderboard
Chinese LLM evaluation covering basic capabilities, Chinese-specific features, and academic proficiency. 20+ models ranked across 10 sub-tasks.
Metric: Total Score. Source: www.cluebenchmarks.com. Saturation forecast: Estimated already saturated. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4 | 70.89 | #438 |
| 2 | Claude 2 | 60.94 | #761 |
| 3 | GPT-3.5 Turbo | 59.79 | #865 |
| 4 | ChatGLM2 6B | 53.85 | #1498 |
| 5 | MiniMax-abab5.5 | 53.06 | #1050 |
| 6 | Llama 2 13B Chat | 34.26 | #1465 |
| 7 | vicuna-13B | 31.7 | #1469 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=superclue · How It Works · Data refreshed daily, snapshot 2026-10-20.