CodeClash: Benchmarking Goal-Oriented Software Engineering

arXiv:2511.00839v2 Announce Type: replace-cross Abstract: Current benchmarks for coding evaluate language models (LMs) on concrete, well-specified tasks such as fixing specific bugs or writing targeted tests. However, human programmers do not spend all day incessantly addressing isolated tasks....