reasoning benchmark
C-Eval
C-Eval is a comprehensive Chinese evaluation suite designed to assess advanced knowledge and reasoning abilities of foundation models in a Chinese context. It comprises 13,948 multiple-choice questions across 52 diverse disciplines spanning humanities, science, and engineering, with four difficulty levels: middle school, high school, college, and professional. The benchmark includes C-Eval Hard, a subset of very challenging subjects requiring advanced reasoning abilities.
Updated Aug 7, 2026
Published models18
Registry coverage18
MetricScore
EvidenceB
C-Eval leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AC | 93.3% | 100.0% | 18 | C | |
| 2 | AC | 93.0% | 94.1% | 18 | C | |
| 3 | MA | 92.5% | 88.2% | 18 | C | |
| 4 | AC | 91.9% | 82.3% | 18 | C | |
| 5 | XI | 91.5% | 76.5% | 18 | C | |
| 6 | AC | 91.4% | 70.6% | 18 | C | |
| 7 | AC | 90.5% | 64.7% | 18 | C | |
| 8 | AC | 90.2% | 58.8% | 18 | C | |
| 9 | AC | 90.0% | 52.9% | 18 | C | |
| 10 | MA | 88.3% | 47.1% | 18 | C | |
| 11 | AC | 88.2% | 41.2% | 18 | C | |
| 12 | DE | 86.5% | 35.3% | 18 | C | |
| 13 | AC | 85.1% | 29.4% | 18 | C | |
| 14 | AC | 83.8% | 23.5% | 18 | C | |
| 15 | AC | 77.2% | 17.6% | 18 | C | |
| 16 | AC | 73.2% | 11.8% | 18 | C | |
| 17 | AC | 50.5% | 5.9% | 18 | C | |
| 18 | BA | 40.7% | 0.0% | 18 | C |
Score distribution
Top published rows on the benchmark's original scale.
C-Eval highlights
The top published results on this benchmark's own scale.
What is C-Eval?
Definition and scoring fields from the benchmark registry.
C-Eval is a comprehensive Chinese evaluation suite designed to assess advanced knowledge and reasoning abilities of foundation models in a Chinese context. It comprises 13,948 multiple-choice questions across 52 diverse disciplines spanning humanities, science, and engineering, with four difficulty levels: middle school, high school, college, and professional. The benchmark includes C-Eval Hard, a subset of very challenging subjects requiring advanced reasoning abilities.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- C-Eval
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- c-eval|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about C-Eval.
Which model scores highest on C-Eval?
Qwen3.6 Plus is currently ranked first with 93.3%.
What does C-Eval measure?
C-Eval is a comprehensive Chinese evaluation suite designed to assess advanced knowledge and reasoning abilities of foundation models in a Chinese context. It comprises 13,948 multiple-choice questions across 52 diverse disciplines spanning humanities, science, and engineering, with four difficulty levels: middle school, high school, college, and professional. The benchmark includes C-Eval Hard, a subset of very challenging subjects requiring advanced reasoning abilities.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
18 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is marked as eligible for the current LLMBoard capability methodology.