mini-SWE-agent vs SWE-agent
A side-by-side look at pricing, capabilities, pros, cons, and our editorial scores.
Tagline
mini-SWE-agent
A 100-line open-source coding agent that scores 74%+ on SWE-bench Verified.SWE-agent
Open-source autonomous agent framework that lets LLMs fix GitHub issues and find security vulnerabilities by using a custom agent-computer interface.Pricing
mini-SWE-agent
FreeΒ· Free and open source (Apache 2.0); you pay only for the underlying LLM tokens.SWE-agent
FreeΒ· Free and open-source; you pay your own LLM API costsAPI
mini-SWE-agent
YesSWE-agent
Not listedPlatforms
mini-SWE-agent
api
SWE-agent
cli
Open source
mini-SWE-agent
Yes Β· MITSWE-agent
Yes Β· MITGitHub stars
mini-SWE-agent
9,575
checked 2026-11-13
SWE-agent
9,575
checked 2026-11-13
Last GitHub push
mini-SWE-agent
2026-11-09SWE-agent
2026-11-09First commit
mini-SWE-agent
2025-06SWE-agent
2025-06Package downloads (30 d)
mini-SWE-agent
14,144,519 Β· pypi mini-swe-agent
checked 2026-11-18
SWE-agent
14,060,333 Β· pypi mini-swe-agent
checked 2026-11-22
Hacker News stories
mini-SWE-agent
2
checked 2026-11-18
SWE-agent
4
checked 2026-11-22
Earliest Wayback capture
mini-SWE-agent
βSWE-agent
2024-11-04
earliest archived capture of the site, not a launch date
Model used
mini-SWE-agent
Multi-model (via litellm, OpenRouter, Portkey)SWE-agent
Multi-model (GPT-4o, Claude Sonnet, DeepSeek, local via LiteLLM)Best for
mini-SWE-agent
Pick mini-SWE-agent if you want a tiny, readable, model-agnostic coding-agent baseline to benchmark, extend, or embed in a research pipeline.SWE-agent
Pick SWE-agent if you're a researcher or engineer who wants a hackable, reproducible coding-agent baseline you can wire to any LLM.Not for
mini-SWE-agent
Skip it if you want a polished IDE assistant or a managed autonomous-engineer SaaS with dashboards and team features.SWE-agent
Skip it if you want a plug-and-play hosted coding assistant - reach for Cursor, Devin, or Cline instead.Editorial score
mini-SWE-agent
7.2 / 10SWE-agent
7.0 / 10Use cases
mini-SWE-agent
swe-bench-evaluationautonomous-codingagent-researchbatch-inferenceagent-baseline
SWE-agent
github-issue-fixingautonomous-codingswe-benchctf-securityagent-research
Pros
mini-SWE-agent
- Roughly 100 lines of Python β trivially readable and hackable
- Scores 74%+ on SWE-bench Verified despite the minimal design
- Model-agnostic via litellm, OpenRouter, and Portkey
- Sandbox-friendly: Docker, Podman, Singularity, Bubblewrap, local
- Apache 2.0; backed by the SWE-bench and SWE-agent authors
SWE-agent
- Open-source under MIT with a strong research pedigree (Princeton/Stanford)
- Model-agnostic via LiteLLM - swap GPT-4o, Claude, or local models freely
- Reproducible SWE-bench harness makes it a credible baseline for agent research
- EnIGMA mode extends the same loop to CTF-style security tasks
Cons
mini-SWE-agent
- No GUI, IDE integration, or hosted product
- Bash-only tool surface β no built-in browser, search, or planning tools
- You build your own ops layer (logging, cost caps, retries)
- Aimed at researchers more than day-to-day end users
SWE-agent
- Officially in maintenance mode; team now recommends mini-swe-agent
- No hosted product, GUI, or managed service - CLI and YAML only
- LLM API costs on long-horizon tasks can be significant
- Setup requires Docker and comfort with Python tooling
Editorial score: rule-based, 0β10, from AI-assisted profile inputs (see /methodology) β not a user rating; βββ means unscored. βNot listedβ means we have no record of it, not that it is absent. GitHub figures and prices carry the date they were checked or captured; prices are shown as published, unconverted.
Pick mini-SWE-agent if
- β Roughly 100 lines of Python β trivially readable and hackable
- β Scores 74%+ on SWE-bench Verified despite the minimal design
- β Model-agnostic via litellm, OpenRouter, and Portkey
- β Sandbox-friendly: Docker, Podman, Singularity, Bubblewrap, local
Pick SWE-agent if
- β Open-source under MIT with a strong research pedigree (Princeton/Stanford)
- β Model-agnostic via LiteLLM - swap GPT-4o, Claude, or local models freely
- β Reproducible SWE-bench harness makes it a credible baseline for agent research
- β EnIGMA mode extends the same loop to CTF-style security tasks