Unverified50% confidenceFactExact time
HumanEval、MBPP、SWE-bench是AI编程评测领域的行业标准基准测试
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
AI编程为何进步最快？四大结构性优势深度解析
bilibili卢菁博士_北大AI博士后6/5/2026
Related Claims
UnverifiedHumanEval、MBPP、SWE-bench是专门针对代码生成的标准评测集83% similarUnverified业界普遍依赖 HumanEval、SWE-bench 等基准测试评估LLM编程能力81% similarUnverified编程能力的评估通常参考HumanEval、MBPP、SWE-bench等标准化基准测试，但实际工程场景表现与基准测试结果存在差异80% similarUnverifiedAI 编程领域最受认可的评测基准包括 HumanEval（测试从函数文档字符串生成正确实现）和 SWE-bench（在真实 GitHub 仓库中定位并修复实际 issue）78% similarUnverifiedHumanEval、SWE-bench等代码评测基准的排名是评估执行型Agent底层模型选型的重要参考维度76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/53528API
curl https://kongchang.com/api/v1/knowledge/claims/53528MCP
get_claim(id=53528)