A Chinese multiple-choice benchmark spanning humanities, social sciences, and STEM subjects to evaluate knowledge and reasoning in Chinese.
Browse the latest scores, model modes, release dates, and parameter sizes for C-Eval.
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
| Rank | Model | License | |||
|---|---|---|---|---|---|
 Qwen3-Max-Thinking Thinking Enabled | 93.70 | 2026-01-26 | 1000B | Closed | |
Qwen 3.6 Plus Preview Thinking Enabled | 93.30 | 2026-03-31 | Unknown | Closed | |
Qwen3.5-397B-A17B Thinking Enabled | 93.00 | 2026-02-16 | 39.7B | Free Commercial | |
4 |  Hunyuan-T1 Standard Mode | 91.80 | 2025-03-21 | Unknown | Closed |
5 | Qwen3.6-27B Thinking Enabled | 91.40 | 2026-04-22 | 27B | Free Commercial |
6 | Qwen3.5-27B Thinking Enabled | 90.50 | 2026-02-25 | 27B | Free Commercial |
7 | Qwen3.6-35B-A3B Thinking Enabled | 90.00 | 2026-04-16 | 35B | Free Commercial |
8 | Qwen3.5-9B Thinking Enabled | 88.20 | 2026-03-02 | 9B | Free Commercial |
9 | Qwen3-32B Thinking Enabled | 87.30 | 2025-04-28 | 32B | Free Commercial |
10 | Qwen3-32B Standard Mode | 83.30 | 2025-04-28 | 32B | Free Commercial |
