← Back to Model Beat
Open Source·12h ago·all news from October 1, 2026

Zero2Repo: Can Coding Agents Build Repositories from Scratch?

Researchers have introduced Zero2Repo, a new benchmark designed to evaluate how effectively AI coding agents can construct entire software repositories from scratch. Unlike previous testing methods that rely on single-language tasks or manual curation, this tool provides a broader framework for assessing agent performance in end-to-end software development.

Covered by 1 source

  • AarXiv CS.AI↗Pei Yang, Tianyu Shi, Yuhang Yao, Wanyi Chen, Tongyun Yang, Dun Pei, Haonan Wang, Pengbin Feng, Guanxu Yu, Jingchun Huang, Zeyu Zhang, Shuhan Sun, Hao Li, Xiang Li, Jie Xiao, Xinyu Wang, Hanxin Chen, Daqi Li, Qi Jia, Hongshan Lin, Zhizhou Gu, Zijun Tian, Weizhi Du, Lynn Ai, Eric Yang12h ago

Related stories

Open SourceAgentXploit: Autonomous Repository-to-Runtime Red-Teaming for AI AgentsSep 28Open SourceSemantic Navigation for Issue Localization in Code RepositorySep 28Open SourceFrom ASR to ASP: Evaluating Prompt Attack Vulnerabilities Against Open-Source LLMsSep 28Open SourceChestPheNoT: Deployable, Auditable Label-Status-Evidence Extraction from Radiology ReportsSep 29