// Cognition
Devin Desktop pricing and models
The editor formerly called Windsurf, from Cognition, with models from every major lab and daily/weekly usage allowances.
All coding toolsPlans checked 2026-10-27
Devin Desktop: Formerly Windsurf: Cognition renamed the editor Devin Desktop on June 2, 2026, and windsurf.com now redirects to devin.ai.
- Free plan
- Yes
- Paid from
- $20/mo
- Own API key
- No
- Open source
- No
Plans
AI code editor with local and cloud agents.
| Plan | Price | What you get |
|---|---|---|
| Free | Free | Light agent quota; limited model availability |
| Pro | $20/mo | Increased quotas, refreshed daily and weekly; full model availability |
| Max | $200/mo | Significantly higher quotas |
| Teamsteam | $40/moper seat + $80/mo team fee | Per-seat quota for each full developer |
Included models and how they score
LiveBench coding and long-horizon agentic scores, strongest coder first, with what the same heavy month of agent work would cost paying per token.
Cognition lists “frontier models from OpenAI, Claude, Gemini and SpaceXAI” plus open-source models and its own SWE-2, without naming versions.
Compare Devin Desktop with
Devin Desktop FAQ
How much does Devin Desktop cost?
Devin Desktop has a free plan (light agent quota; limited model availability). Paid plans start at $20/mo (Pro). Prices from Cognition’s pricing page, checked 2026-10-27.
Which models does Devin Desktop use?
Cognition lists “frontier models from OpenAI, Claude, Gemini and SpaceXAI” plus open-source models and its own SWE-2, without naming versions.
What changed with Devin Desktop recently?
Formerly Windsurf: Cognition renamed the editor Devin Desktop on June 2, 2026, and windsurf.com now redirects to devin.ai.
Where these numbers come from
- Plans and prices — each vendor's own pricing page, checked 2026-10-27: Devin Desktop. Usage allowances change often; check before you buy.
- Model scores — LiveBench coding and long-horizon agentic categories for the models each plan includes, live from the same data as the leaderboard. A tool's own agent loop matters too; these scores rate the model, not the tool.
- “Same work via API” — a heavy coding-agent month (40K tokens in, 4K out, 20K requests, 70% cached input) at OpenRouter list price, for comparing a subscription against paying per token.