开发

Why we no longer evaluate SWE-bench Verified

归档编码:why-we-no-longer-evaluate-swe-bench-verified-33e5e5bc

核心要点速览

SWE-bench Verified is increasingly contaminated and mismeasures frontier coding progress. Our analysis shows flawed tests and training leakage. We recommend SWE-bench Pro.

首发信源 OpenAI News OpenAI 官方博客、模型发布与科研进展
所属板块 开发 AIPEDIAS 实时情报监测
查看完整报道与分享

完整图文、背景资料与讨论归原始发布平台 OpenAI News 所有。

直达原文报道 ↗