reasoning benchmark
HumanEval
A benchmark that measures functional correctness for synthesizing programs from docstrings, consisting of 164 original programming problems assessing language comprehension, algorithms, and simple mathematics
Updated Aug 7, 2026
Published models66
Registry coverage66
MetricScore
EvidenceB
HumanEval leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
Score distribution
Top published rows on the benchmark's original scale.
HumanEval highlights
The top published results on this benchmark's own scale.
What is HumanEval?
Definition and scoring fields from the benchmark registry.
A benchmark that measures functional correctness for synthesizing programs from docstrings, consisting of 164 original programming problems assessing language comprehension, algorithms, and simple mathematics
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- HumanEval
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- humaneval|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about HumanEval.
Which model scores highest on HumanEval?
MiniCPM-SALA is currently ranked first with 95.1%.
What does HumanEval measure?
A benchmark that measures functional correctness for synthesizing programs from docstrings, consisting of 164 original programming problems assessing language comprehension, algorithms, and simple mathematics
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
66 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.