LLM Stats (C-Eval): leaderboard
13,948 multi-choice questions across 52 disciplines spanning STEM, humanities, social science, and professional fields, evaluating Chinese knowledge.
Metric: Score (%). Source: llm-stats.com. Saturation forecast: Estimated already saturated. 20 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 Max (Thinking) | 93.7 | #208 (Qwen 3 Max) |
| 2 | Qwen 3.6 Plus | 93.3 | #120 |
| 3 | Qwen 3.5 397B A17B | 93 | #146 |
| 4 | Qwen 3.5 122B A10B | 91.9 | #173 |
| 5 | MiMo-V2.5-Pro | 91.5 | #158 |
| 6 | Qwen 3.6 27B | 91.4 | #170 |
| 7 | Qwen 3.5 27B | 90.5 | #187 |
| 8 | Qwen 3.5 35B A3B | 90.2 | #256 |
| 9 | Qwen 3.6 35B A3B | 90 | #226 |
| 10 | Qwen 3.5 9B | 88.2 | #379 |
| 11 | DeepSeek V3 | 86.5 | #321 |
| 12 | Qwen 3 14B | 86.2 | #537 |
| 13 | Qwen 3.5 4B | 85.1 | #475 |
| 14 | Qwen 2 72B Instruct | 83.8 | #576 |
| 15 | Qwen 2 7B Instruct | 77.2 | #991 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-c-eval · How It Works · Data refreshed daily, snapshot 2026-10-20.