repobench is an open-source Python package for running repository-native evaluations of coding agents. It provides benchmark and evaluation workflows for assessing agent performance on software engineering tasks in realistic code repositories.
In the Agent evaluation & testing space, repobench takes a focused approach. It focuses on evaluating coding agents against realistic repository-based software engineering tasks. repobench is an open-source project aimed at AI researchers and developers building coding agents. The project is open source (Apache-2.0). repobench is available on the command line, and it can be self-hosted.
It is developed by AgentFit, and it first shipped in 2026. Key capabilities include repository-native evals, coding-agent benchmarks, and software engineering tasks.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Same category — not a similarity match