Windsurf (now Devin Desktop) pricing guide details
Pricing Overview
Windsurf is now sold as Devin Desktop, part of Cognition's Devin platform, and it is priced on Devin's plans rather than its own. Cognition announced its agreement to acquire Windsurf in July 2025; windsurf.com now redirects to devin.ai/desktop, and the Windsurf name appears nowhere on the current pricing page.
There are three individual plans and two team plans. Free is $0, Pro is $20 per month and Max is $200 per month. The Team plan is charged in two parts: $80 per month for the team itself, plus $40 per month for each full developer seat. Enterprise is quote-only.
The pricing model is quota-based rather than credit-based. Every paid plan carries a usage allowance that refreshes on both a daily and a weekly cycle, and Cognition states that cost per message varies with the model, the task size and the reasoning required. If you exceed the allowance on a paid plan you can buy extra usage, which is consumed at API pricing.
Plan Comparison
| Plan | Price | Members | Concurrent sessions |
|---|---|---|---|
| Free | $0 | 1 | Limited |
| Pro | $20/month | 1 | Up to 10 |
| Max | $200/month | 1 | Up to 10 |
| Team | $80/month + $40/month per full seat | Unlimited flex seats | Unlimited |
| Enterprise | Custom | Unlimited | Unlimited |
The tiers are cumulative: Pro is everything in Free, Max is Pro with significantly higher quotas, Team is Pro for everyone, and Enterprise builds on Team.
Free is limited in which models it can reach as well as in volume. Pro is where full model availability begins, including OpenAI, Claude and Gemini frontier models, free use of Cognition's SWE 1.7 model, and access to cloud agents through Devin Cloud. Max buys quota rather than capability. Team adds sharing, centralised billing, an admin dashboard with analytics and priority support, and each full seat carries its own quota. Enterprise adds SAML/OIDC SSO, centralised admin controls, deployment in your own VPC, teamspace isolation and custom terms.
Integration coverage widens with the plan: Slack and Teams, Linear and Jira, and GitHub, GitLab and Bitbucket appear on the paid tiers, with custom git provider support reserved for Enterprise.
Hidden Costs and Considerations
The team platform fee is separate from seats. Team is $80 per month before any seat is counted, then $40 per full developer seat. A five-developer team is $280 per month, not $200. This is the detail most easily missed when comparing against per-seat-only competitors.
Usage is metered above the allowance. The quota refreshes daily and weekly, and overage is bought separately and consumed at API pricing, so heavy agent use raises the bill above the plan price. Cognition does not publish a per-message rate, stating instead that it depends on the model, task size and reasoning required.
Free and flex seats are not the same thing. Team allows unlimited members on flex seats, but only a full seat at $40 per month carries its own quota and Devin Desktop access. Headcount is therefore not the same as billable seats.
No annual pricing is published. Every figure on Cognition's page is monthly.
Cost Estimates by Team Size
The figures below apply the published Team formula — $80 for the team plus $40 per full seat. They are arithmetic on Cognition's own prices, not quotes.
| Team | Plan | Monthly |
|---|---|---|
| 1 developer | Pro | $20 |
| 1 heavy user | Max | $200 |
| 5 full seats | Team | $280 |
| 20 full seats | Team | $880 |
| 50 full seats | Team | $2,080 |
Add model usage above the included allowance to every paid row. An organisation needing SSO, VPC deployment or teamspace isolation should price Enterprise instead, since those controls are not available on Team at any seat count.