HumanEval

OpenAI · v2024

Details

Purpose
Measure functional correctness of code generation from docstrings
Domain
Code generation
Health
Saturating
Reproducibility
Reproducible
Contamination concern
High
Saturation
Saturated for frontier models
Code available
Yes
Last reviewed
2026-06-15

Benchmark run records and model comparisons will be added in Phase 4.