Unverified50% confidenceBenchmarkExact time
HumanEval专注代码生成能力，以164道编程题检验模型的逻辑推理与代码正确性
1
Sources
50%
Confidence
Long-term
Relevance
7/9/2026
First Seen
Sources
大模型评测：测试行业最稀缺的转型方向与入场时机
bilibili宇宙Ai测试7/6/2026
Related Claims
VerifiedHumanEval基准包含164道编程题，是衡量代码生成能力的主流标准83% similarUnverified代码大语言模型采用HumanEval、MBPP等可执行基准评估，即直接运行生成代码并检验输出是否正确77% similarUnverifiedHumanEval测试代码生成能力，要求模型根据函数签名和文档字符串生成正确的Python代码76% similarUnverified代码生成领域的模型评估如HumanEval、SWE-bench格外强调功能正确性而非流畅度76% similarUnverifiedHumanEval由OpenAI自研，专门测试代码生成的功能正确率73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/389905API
curl https://kongchang.com/api/v1/knowledge/claims/389905MCP
get_claim(id=389905)