Sources

4 Total Sources

OpenAI Finds 30% of SWE-Bench Pro Tasks Broken, Retracting Earlier Recommendation

OpenAI11h ago
OpenAI Audit Reveals ~30% of SWE-Bench Pro Coding Benchmark Tasks Are Broken
ai.cm21h ago
DeepSWE Benchmark Resets the AI Coding Race, Crowning GPT-5.5
arxiv.org11h ago
SWE Atlas: Benchmarking Coding Agents Beyond Issue Resolution
github.com11h ago
GitHub - benchflow-ai/awesome-evals: A curated, non-BS library of the best resources for building and evaluating AI agents — papers, blogs, talks, tools, benchmarks. Maintained by BenchFlow. · GitHub