Capability
Benchmark scores · average 66.6 · category avg 61.0
Momentum
30-day star trajectory · velocity 49/100
The Stack
Brain · Harness · Tools · Environment
Frequently asked
Pulled from the dataset · updated daily
What is SWE-agent?
SWE-agent is the research agent from the Princeton NLP group behind SWE-bench. Introduced the agent-computer interface (ACI) idea · a restricted terminal-like API optimized for LLM tool use.
What model powers SWE-agent?
SWE-agent runs on claude-sonnet-4-6 and gpt-5. It uses a react loop pattern with custom tools.
Is SWE-agent open-source?
Yes · SWE-agent is fully open-source under the MIT license. The harness, tools, and loop code are all public on GitHub at SWE-agent/SWE-agent.
How popular is SWE-agent?
SWE-agent has 18,962 stars on GitHub and gained 65 in the last 7 days across 96 contributors.
How does SWE-agent compare to other autonomous agents?
SWE-agent scores 66.6 on the composite capability rank · 5.6 points above the autonomous category average (61.0) across 4 peer agents.
Brains used
GitHub stats
Composite capability · momentum formula · data sources.
Also in Autonomous
Peer agents in the same category · sorted by capability