In 2024, we built SWE-bench
and SWE-agent at Princeton University
and helped kickstart the coding agent revolution.
These works have influenced most modern AI agents, including Claude Code, GitHub copilot and many others.
Our latest follow-up work is the mini agent:
Performant: Scores >74% on the SWE-bench verified benchmark;
starts faster than Claude Code
Deployable: In addition to local envs, you can use docker, podman,
singularity, apptainer, and more