SWE-agent
Open-source AI agent that autonomously fixes GitHub issues using your choice of LLM
SWE-agent is a landmark research tool, but its maintenance mode makes it a hard sell for new projects. If you need EnIGMA's CTF capabilities or want to study the original architecture, it's worth the setup. For everything else, mini-swe-agent is the smarter default — lighter, simpler, and actively developed. If you need a supported, polished commercial alternative, GitHub Copilot Workspace or Devin are better fits but cost money and offer less transparency.
Verified 6d ago · liveness 74/100 · cite: rightaichoice.com/tools/swe-agent
- Researchers studying autonomous program repair or agent-computer interfaces
- Security researchers automating penetration testing with EnIGMA mode
- Developers comfortable with CLI and Docker who want to automate bug triage
- Hobbyists experimenting with AI-driven code generation
- Beginners unfamiliar with command-line tools or Docker
- Users needing a visual interface or IDE plugin
- Teams without access to LLM API keys or local models
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip SWE-agent if you need an actively maintained, production-supported tool, prefer a visual interface, or are not comfortable with Docker and command-line setup — in those cases mini-swe-agent or a commercial alternative like GitHub Copilot Workspace is a better fit.
Running SWE-agent yourself means you pay for your own LLM API usage (e.g., GPT-4o or Claude Sonnet 4 tokens), which can add up quickly if you process many issues.
SWE-agent is completely free (MIT license), but you pay for your own LLM API tokens and infrastructure. For researchers and hobbyists, it's cheaper than commercial tools like GitHub Copilot Workspace or Devin, which charge monthly subscription fees. However, commercial tools bundle support and a polished UI, so the trade-off is control and transparency versus convenience. If you're cost-sensitive and comfortable with CLI and Docker, SWE-agent is the most affordable option; otherwise, budget for
In short
SWE-agent — Open-source AI agent that autonomously fixes GitHub issues using your choice of LLM. Best for Researchers studying autonomous program repair or agent-computer interfaces, Security researchers automating penetration testing with EnIGMA mode, Developers comfortable with CLI and Docker who want to automate bug triage. Free to use.
What's new in SWE-agent
Checked 6 days agoAcross the latest 4 updates: 4 feature updates.
Releasing SWE-agent 1.0: SoTA on SWE-bench light & tons of new features
SWE-agent 1.0 achieved state-of-the-art results on SWE-bench light and introduced numerous new features for autonomous code fixing.
SWE-agent-LM-32b achieves open-weights SOTA on SWE-bench
The team's own 32b language model achieved state-of-the-art results among open-weights models on SWE-bench.
SWE-agent 1.0 + Claude 3.7 is SoTA on SWE-Bench full
Combining SWE-agent 1.0 with Claude 3.7 achieved state-of-the-art performance on the full SWE-bench benchmark.
SWE-agent 1.0 + Claude 3.7 is SoTA on SWE-bench verified
SWE-agent 1.0 paired with Claude 3.7 achieved state-of-the-art results on the verified subset of SWE-bench.
What people actually say about SWE-agent — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
30 mentions across 1 source (Hacker News) · researched Jul 3, 2026.
- +Autonomous bug fixing from a GitHub issue URL.
- +Supports multiple LLM backends: OpenAI, Anthropic, local models.
- +Sandboxed Docker environment ensures safe execution.
- +Top SWE-bench scores at time of NeurIPS 2024 publication.
- +Free and open-source under MIT license.
- −Replaced by mini-swe-agent for simpler use cases.
- −Complex Docker-based setup and configuration.
- −No GUI; entirely command-line driven.
- −Documentation can feel academic, not user-friendly.
- −Autonomous mode may produce incorrect fixes without review.
- • LLM API usage costs (pay-per-token) not included
- • Compute and storage for Docker images
Viability Score
How well maintained and how widely used is SWE-agent? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: August 2026
How we score →Key Features
- Autonomous GitHub issue fixing
- Use any LLM backend (GPT-4o, Claude Sonnet 4, Ollama)
- Local models via Ollama
- Docker sandbox for safe execution
- Automatic pull request creation
- Single YAML configuration
- Trajectory logging for debugging
- EnIGMA mode for cybersecurity CTF challenges
- SWE-bench benchmarking
- Command-line interface
- GitHub API integration
- Extensible tool set (file editing, shell)
- Competitive coding challenges support
- MIT license
- Open source from Princeton and Stanford
About SWE-agent
SWE-agent is an open-source, MIT-licensed toolkit from Princeton and Stanford researchers that lets you point a language model of your choice at a GitHub issue and have it autonomously work toward a fix. It equips the model with tools to explore a repository, edit files, and execute shell commands inside a Docker sandbox, then generates a pull request with the proposed changes. Built for developers, researchers, and security engineers, it's also been used for offensive cybersecurity challenges and competitive coding. The project recently hit state-of-the-art results on SWE-bench among open-source projects and was presented at NeurIPS 2024. SWE-agent enables your language model of choice (e.g. GPT-4o or Claude Sonnet 4) to autonomously use tools to fix issues in real GitHub repositories, find cybersecurity vulnerabilities, or perform any custom task. The system is highly configurable — a single YAML file governs the agent's behavior, and full documentation covers installation, benchmarking, and customization. It supports any LLM backend, including GPT-4o, Claude Sonnet 4, and local models via Ollama, and it logs trajectories to help you debug agent reasoning. For security work, the EnIGMA mode tackles capture-the-flag challenges and has achieved top scores on multiple cybersecurity benchmarks. Because it's designed for research, the codebase is simple and extensible, making it a favorite for those studying AI-driven software engineering. That said, the development team has explicitly shifted focus to mini-swe-agent, a lighter version that matches SWE-agent's performance with far less complexity. The original remains functional but is now in maintenance mode. For new users, the maintainers recommend mini-swe-agent. SWE-agent itself offers maximum control and research flexibility, but it demands comfort with the command line and Docker. For teams seeking a polished, supported commercial tool, options like GitHub Copilot Workspace or Devin may be a better fit — though you'll pay for that polish with far less transparency and customizability.
Behind the Verdict
SWE-agent is a powerful research artifact from Princeton and Stanford that demonstrated the potential of agent-computer interfaces for automated software engineering. Its core value proposition is flexibility: you bring your own LLM (GPT-4o, Claude Sonnet 4, or any Ollama-hosted local model) and SWE-agent provides the scaffolding to turn that model into an autonomous coding agent. The single YAML configuration is a standout feature, giving researchers complete control over the agent's behavior without touching the Python source. The trajectory logging is invaluable for debugging why an agent made a particular decision, and the EnIGMA mode for offensive cybersecurity is a unique capability that's hard to find elsewhere. The biggest strengths are the SWE-bench state-of-the-art results and the clean, hackable codebase. It's genuinely made for research: you can tweak the tool definitions, experiment with different prompts, and measure performance against a standardized benchmark. The EnIGMA mode has achieved top scores on multiple cybersecurity leaderboards, making it a go-to tool for security researchers exploring AI-driven penetration testing. However, the project's development status is a major consideration. The maintainers have explicitly stated that most development effort is now on mini-swe-agent, a lighter version that matches SWE-agent's performance with far less complexity. The original repository is in maintenance mode — it still works, but you shouldn't expect new features. This is fine for research where reproducibility matters, but risky for production use. There's also a learning curve: you need to be comfortable with the command line, Docker, and YAML configuration. If you want a visual interface or an IDE plugin, this isn't it. Weak LLMs will produce incorrect patches, so the tool's effectiveness is gated by the quality of the model you pair it with. For a researcher studying agent design or a security engineer wanting to benchmark AI penetration testing, SWE-agent is still a valuable resource. For most other use cases, mini-swe-agent is the better choice. If you need a polished commercial tool, GitHub Copilot Workspace or Devin offer more hand-holding but cost money and give you less control.
Researching SWE-agent? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas SWE-agent actually fits — and what changes day-one when you adopt it.
You want to evaluate how well Claude Sonnet 4 fixes bugs in a set of GitHub repositories.
Outcome: Using SWE-agent with a YAML configuration targeting Claude Sonnet 4, you run the agent on a SWE-bench subset, collect trajectories, and analyze the success rate, publishing results in a paper.
You need to automate solving capture-the-flag challenges for a security assessment.
Outcome: Using EnIGMA mode (SWE-agent 0.7), you configure the agent for CTF tasks, run it against a challenge set, and achieve state-of-the-art scores, saving hours of manual exploitation work.
Your team has a backlog of open GitHub issues, and you want to generate candidate fixes automatically.
Outcome: You set up SWE-agent with your preferred LLM (e.g., GPT-4o), point it at priority issues, and it generates pull requests with proposed changes, which your team reviews and merges or refines.
Use Cases
- Automatically generate pull requests for open GitHub issues
- Evaluate LLM performance on software engineering tasks using SWE-bench
- Simulate autonomous agents for cybersecurity red team exercises
- Practice competitive programming with AI co-pilots
- Prototype automated bug bounty solvers
- Conduct research on code generation and repair
Models Under the Hood
as of 2026-08-21
Limitations
- SWE-agent requires running locally or on a server with Docker.
- Performance depends heavily on the underlying LLM; weak models may produce incorrect patches.
- The tool is best for straightforward bugs and may struggle with highly complex or context-dependent issues.
- The original tool is now in maintenance mode; most development effort has shifted to mini-swe-agent.
as of 2026-08-19
Verification history
We have re-verified SWE-agent 6 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published SWE-agent tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0
Ideal for
Researchers, hobbyists, and security engineers who want a free, open-source tool to experiment with autonomous code fixing or cybersecurity CTF challenges, provided they're comfortable with CLI and Docker.
What this tier adds
Starting tier: MIT-licensed, free forever, with full features including EnIGMA mode, but you pay for your own LLM API tokens and infrastructure.
Where the pricing makes sense
The company stage and team size where SWE-agent's pricing actually pencils out — and where peers do it cheaper.
SWE-agent is completely free (MIT license), but you pay for your own LLM API tokens and infrastructure. For researchers and hobbyists, it's cheaper than commercial tools like GitHub Copilot Workspace or Devin, which charge monthly subscription fees. However, commercial tools bundle support and a polished UI, so the trade-off is control and transparency versus convenience. If you're cost-sensitive and comfortable with CLI and Docker, SWE-agent is the most affordable option; otherwise, budget for
Setup time & first value
How long it actually takes to get something useful out of SWE-agent — broken out by persona, not the marketing-page minute.
For a developer familiar with Docker and Python: roughly 1-2 hours to install via pip, configure a YAML file, and run a first simple issue fix. For a researcher wanting to run SWE-bench benchmarks: half a day to a day to set up the full evaluation pipeline. For a security user wanting EnIGMA mode: expect a few hours to install the specific 0.7 version and configure the CTF environment.
Switching to or from SWE-agent
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From manual GitHub issue fixing: run SWE-agent on a test issue first to see if the generated patch matches your coding style, then adopt for routine bug triage.
- →From mini-swe-agent: if you need EnIGMA's cybersecurity mode or want to experiment with the original architecture, install SWE-agent 1.0 and port your mini-swe-agent configuration.
- ↗To mini-swe-agent: if you find SWE-agent too complex, mini-swe-agent matches its performance with a simpler setup; port your YAML configuration and testing scripts.
- ↗To GitHub Copilot Workspace: if you prefer a commercial, supported tool with a UI, export your GitHub issue backlog and start using Copilot Workspace's guided workflow.
- ↗To Devin: for a fully managed autonomous coding assistant, migrate your GitHub repository and issue tracker to Devin's platform.
Integrations
Resources & Guides
Tutorials & Learning
Official links
Tools that pair well with SWE-agent
Common stack mates teams adopt alongside SWE-agent, with the specific reason each pairing earns its keep.
OpenHands
Open-source platform for autonomous cloud coding agents that fix bugs, review PRs, and automate workflows.
Open Interpreter
Open-source terminal agent that runs natural-language commands on your computer
Refact.ai
Open-source autonomous AI coding agent that plans, executes, and deploys tasks in your IDE.
Featured Head-to-Head Comparisons
Swe Agent vs Locus Robotics
Locus Robotics and SWE-agent solve completely different problems: warehouse logistics vs. software bugs. Your choice depends on domain. If you need to physically move goods in a warehouse, Locus is proven with real-world clients; if you need to automate GitHub issue resolution, SWE-agent is free and open-source. There is no overlap—buyers should evaluate based on their operational need, not feature comparison.
Swe Agent vs Truleo
If you're in law enforcement needing to unearth leads from scattered data, Truleo's purpose-built suite (jail call analysis, BWC analysis, report writing) is unmatched. For developers automating bug fixes, SWE-agent's open-source flexibility and LLM-agnostic design make it a powerful free tool. Choose based on your domain: policing or programming.
Swe Agent vs Presto Voice
These tools serve completely different markets: Presto Voice is a commercial drive-thru automation platform for QSR chains, while SWE-agent is a free, open-source bug-fixing tool for developers. Choose Presto if you run a multi-location QSR and want to boost revenue via voice AI; choose SWE-agent if you manage open-source projects and need automated patch generation. There is no overlap.
Alternatives to SWE-agent
View allOpenHands
Open-source platform for autonomous cloud coding agents that fix bugs, review PRs, and automate workflows.
Open Interpreter
Open-source terminal agent that runs natural-language commands on your computer
Frequently Asked Questions
Categories
Best-of guides
Used SWE-agent? Help shape our editorial sentiment research.