SWE-Agent
AI agent that autonomously fixes GitHub issues
AI Agent FrameworksFree (open-source)★ 15,000Works with OpenClaw
About SWE-Agent
SWE-Agent is an AI agent built by Princeton that can autonomously resolve GitHub issues. It navigates codebases, understands context, makes edits, and submits pull requests — achieving strong results on the SWE-bench benchmark.
Features
Issue fixing
GitHub integration
Autonomous debugging
Code search
Multi-model
The tally
FOR
- +Strong benchmark performance
- +GitHub-native workflow
- +Open-source and research-backed
- +Multi-model support
- +Active development
AGAINST
- −Research-oriented (not polished)
- −Requires significant compute
- −Can be slow per issue
- −Limited to GitHub workflow
Related concepts
Kept nearby
Abacus AI DeepAgent
AI agent that builds full-stack apps and automations from natural language
freemium
Microsoft AutoGen
Framework for building multi-agent conversational AI
Free (open-source) · ★ 35,000
Superagentic
Full-stack agentic AI company — tools and consulting for production agents
Open-source tools + enterprise consulting · ★ 2,000
Smolagents
Hugging Face's lightweight library for building agents
Free (open-source) · ★ 15,000
Browse all AI Agent Frameworks tools →
Featured in
AI Coding
Coding-Agent Benchmark Methodology Checklist
AI Tools
ChatGPT vs Claude vs Gemini for Coding in 2026: Which AI Wins?
Tools & APIs
Best Vibe Coding Tools in 2026: AI Assistants That Keep You in Flow State
Tools & APIs
GitHub Copilot vs Tabnine vs Amazon Q vs Gemini Code Assist: Best AI Coding Assistant for Teams in 2026