AUTOMATIONSWITCH
Home/AI Coding Assistants/OpenHands
Open Source · Autonomous · All-Hands-AI

OpenHands

The best-funded open-source autonomous coding agent ($18.8M Series A). MIT license, 71K+ GitHub stars, 53-77% SWE-bench.

Editorial Score
3/5
Visit OpenHands →GitHub (71.2K) →← All Tools
3/5
Editorial score
MCP SUPPORTED
Mode
Autonomous
IDE Support
Web, CLI
License
Open Source
Pricing
Free
MCP Support
Yes
Vendor
All-Hands-AI
Free (open source, MIT license, BYOK)
// Key Numbers
70K+
GitHub stars
490+
Contributors
$18.8M
Series A
MIT
License
~77%
SWE-bench Verified (best)
// Our Verdict

The best-funded open-source autonomous agent. Strong SWE-bench scores and MIT license make it the open-source alternative to Devin.

Best for: Teams who want autonomous AI coding agents with full control, self-hosting, and no vendor lock-in.
// Pricing
Open Source (self-hosted)
Free
MIT-licensed core, Docker sandbox, full agentic capabilities, BYOK model support
Cloud Individual
Free
Managed SaaS, limited daily conversations and parallel runs, OpenHands LLM provider at cost
Cloud Pro
$20/mo
Runtime compute included, BYOK model keys (OpenAI, Anthropic, Mistral), higher usage limits
Enterprise
Custom
Self-hosted in your VPC via Kubernetes, multi-user support, centralised access management, priority support
// Our Assessment

OpenHands (formerly OpenDevin) is the most popular self-hosted autonomous coding agent, with 70K+ GitHub stars, 490+ contributors, and an $18.8M Series A led by Madrona. It writes code, runs commands, browses the web, and opens PRs, all inside a sandboxed Docker environment. Model-agnostic by design, it works with Claude, GPT-4o, Gemini, or local models via OpenRouter. On SWE-bench Verified it scores 53-77% depending on the underlying model.

For teams that cannot send code to third-party clouds, OpenHands is the answer. The MIT-licensed core runs on your infrastructure via Docker or Kubernetes (v1.6.0, March 2026). OpenHands Cloud offers managed SaaS with a free individual tier and paid team plans.

The trade-off is setup complexity. Docker, API keys, and configuration are non-trivial compared to downloading Cursor. But for organisations needing code sovereignty and model flexibility, OpenHands delivers what no proprietary tool can.

// Ideal For
+Engineering teams in regulated industries that need code sovereignty, self-host on your own infrastructure with MIT license
+Organisations that want model flexibility without vendor lock-in, works with any LLM backend
+Open-source contributors who want to extend and customise their coding agent (490+ contributors, active SDK)
+Teams evaluating Devin-class autonomous agents without the $500/mo price tag
// Not Ideal For
–Developers who want zero-setup tooling, Docker configuration and API key management add friction vs Cursor or Copilot
–Solo developers seeking real-time autocomplete and inline suggestions (OpenHands is task-oriented, not keystroke-level)
–Teams without DevOps capacity to maintain self-hosted infrastructure
// Strengths
+Best-funded open-source coding agent ($18.8M Series A)
+MIT license, fully open and self-hostable
+53-77% SWE-bench Verified, competitive with proprietary agents
+Active development with strong contributor community
// Weaknesses
–Autonomous agents require significant trust and well-scoped tasks
–Self-hosting requires infrastructure and maintenance
–Less polished UX than commercial alternatives
–Resource-intensive, requires significant compute
// Models Supported
ClaudeGPT-4oGeminiAny OpenAI-compatible
// Related Comparisons
Devin vs OpenHands vs Copilot Agent →
Proprietary vs open-source autonomous agents, cost, control, and capability compared
Claude Code vs Codex CLI vs Gemini CLI →
How OpenHands' model-agnostic approach compares to vendor-locked CLI agents
Agentic vs Autocomplete AI Coding →
Where OpenHands' autonomous task execution fits in the coding AI spectrum
Page Feedback

Spot something stale, broken, or unclear?

Send a correction or note. We review submissions privately before changing the page.

Human submissions are rate-limited and forwarded to the editorial inbox. Agent submissions can use the same endpoint with an authenticated key.

// AI Coding Audit
NOT SURE IF OPENHANDS
FITS YOUR WORKFLOW?

We map your development workflow, evaluate which AI coding tools fit your team size, stack, and security requirements, and give you a prioritised adoption plan.

Book an AI Coding Audit →Or visit OpenHands directly →
← Back to all AI Coding Assistants