people · May 16, 2026

SWE-agent Scores Up to 74% on SWE-bench Verified in 100 Lines of Python Code

The SWE-bench Leaderboards site updated mini-SWE-agent scores to 74% on SWE-bench Verified. Mini-SWE-agent is the 100-line AI agent built by the Princeton and Stanford team behind SWE-bench and SWE-agent. The site lists mini-SWE-agent as widely adopted by Meta, NVIDIA, Essential AI, IBM, Nebius, Anyscale, Princeton University, and Stanford University. The Princeton and Stanford team released mini-SWE-agent as a minimal version of the original SWE-agent framework.