LetTokenBurn
Sign in
Verified100%Harbor adapterCodingMultimodal

HumanEval

Human Evaluation

HumanEval: 164个函数生成

PaperRepoDatasetLeaderboardHomepage

How to run

ltb · humaneval
$ npm i -g lettokenburn-cli
$ ltb bench pull humaneval --n 50
$ ltb bench run ./humaneval -a claude-code -m anthropic/claude-opus-4-8
$ ltb bench push

Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs

Scores

4 sourced entries
HarnessSourceConfidenceLink
Llama399.4N/AIndependentmediumFeb 25, 2024arxiv.org
Phi-298.2N/AIndependentmediumNov 21, 2024arxiv.org
Mistral 7B93.9N/AIndependentmediumMay 18, 2024arxiv.org
Claude 3.5 Sonnet90.8N/AIndependentmediumN/Apaperswithcode.com

Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.

Reproduced by the community

All runs
Loading community runs…

About

HumanEval (Human Evaluation) — 164个函数生成. Scoring: Pass@1. Difficulty: 低（饱和95%+）. Imported from benchmark_research.md (2026-06).