Reproduce “CrossSWE: Benchmarking Cross-Repository Joint Development for Coding Agents” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Silin Chen, Yuling Shi, Heng Lian, Zihan Zhang, Xiaodong Gu, Haotian Lin, Zhensu Sun
CrossSWE tests whether tasks require coordinated changes across repositories by checking that each task fails with any partial set of repository patches and passes only when all patches are applied. In its 171-task benchmark, both repositories remain available in a single-side comparison, yet GPT-6.1 Sol’s Pass@1 rises from 58.48% to 64.55%, showing added difficulty from joint edits in this setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.