Verified90% confidenceFactExact time
HumanEval由OpenAI于2021年提出，考察函数级代码补全
9
Sources
90%
Confidence
Long-term
Relevance
7/9/2026
First Seen
Sources
GPT-5.6三模型详解：Sol、Terra、Luna分层定位与能力对比
bilibiliX信息差6/28/2026
Related Claims
Cite This Claim
Stable URI
https://kongchang.com/claim/389031API
curl https://kongchang.com/api/v1/knowledge/claims/389031MCP
get_claim(id=389031)