{"slug":"mle-bench","name":"mle-bench","category":"eval","categoryName":"Evaluation, red-teaming & guardrails","audience":["ai-era"],"primary":"ai-era","openClosed":"open","pricing":"open-source","license":"MIT","maintained":"stale","officialUrl":"https://openai.com/index/mle-bench/","repo":"https://github.com/openai/mle-bench","latestVersion":"","latestRelease":"","stars":1769,"starsAsOf":"2026-10-09","description":"MLE-bench is a benchmark for measuring how well AI agents perform at machine learning engineering","descriptionZh":"机器学习工程 agent 基准","callMethods":["CLI"],"platforms":[],"addedAt":"2026-10-10","updatedAt":"2026-10-09","topics":[],"language":"Python","lastPush":"2026-04-24","formerly":null,"packages":[],"recentVersions":[],"alternatives":["promptfoo","openai-evals","deepeval","ragas","lm-evaluation-harness","opencompass","swe-bench","giskard-oss"],"stacks":[],"scenarios":["evaluate-and-monitor"],"problems":[],"reasonZh":"开源大库收录：机器学习工程 agent 基准","sources":["oss"],"record":"https://aimap.page/api/v1/software/mle-bench.json","pages":{"en":"https://aimap.page/en/software/mle-bench/","zh-hans":"https://aimap.page/zh-hans/software/mle-bench/"}}