Unverified50% confidenceBenchmarkExact time
2025年的顶尖模型在HumanEval基准测试中的通过率已突破90%
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/20/2026
First Seen
Valid until: 10/18/2026
Sources
Related Claims
UnverifiedHumanEval等标准化代码基准测试显示，顶级模型的pass@1得分已超过90%80% similarUnverified在HumanEval等代码能力基准测试中，顶级大语言模型的通过率已超过90%76% similarVerified代码生成模型在HumanEval等基准测试中的通过率从60%跃升至90%以上（2024-2025年间）76% similarUnverified当前顶级模型的单步准确率约在90%至95%之间76% similarUnverified自2023年GPT-4发布以来，AI在HumanEval等主流编程基准测试上的通过率从不足50%跃升至90%以上76% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/575444API
curl https://kongchang.com/api/v1/knowledge/claims/575444MCP
get_claim(id=575444)