Skip to main content
In March 2026, Windsurf replaced the credit-based system with a quota-based usage system. Instead of buying and spending credits, your plan now includes a daily and weekly usage allowance that refreshes automatically.

​
How quotas work

Your plan includes a usage allowance measured as a daily and weekly budget. Your budget is based on how many tokens the model uses for each request. The cost per token varies by model, and free models don’t count against your quota at all. Short requests, with only a few files in context, will use fewer tokens than longer requests with larger codebases. This system is different from the previous credit-based system, but better reflects the underlying costs of using different models.

​
When you hit your limit

  • Free: Wait until your next daily or weekly reset.
  • Pro, Teams, or Max: Purchase extra usage to keep working without interruption.
Your quota resets on a daily and weekly basis, based on the calendar date. Your daily quota is more than 1/7 of your weekly quota, enabling users who work on weekends to fully use their weekly allowance.

​
Checking your remaining quota

You can check your remaining quota and when it resets from the usage meter in Windsurf, or on your plan page.

​
Making your quota last longer

  • Be precise with your instructions and remove unnecessary context.
  • Switch to free models like SWE-1.5 for routine tasks.
  • Avoid unnecessarily long sessions when a quick prompt will do.
  • Try to choose a single frontier model for your tasks—requests to the same model leverage caching and reduce overall token usage.

​
Extra usage

Extra usage lets you continue using Windsurf after hitting your included quota. Usage is billed at API list prices for the model you’re using, based on how many tokens the model uses for each request. Priority and speed configurations (e.g., SWE-1.5 Fast, fast Opus variants) will increase the cost.
Quota limits never limit your extra usage, just the built-in allowance from your plan.

​
Migration for existing subscribers

Your price is grandfathered in at $15/mo indefinitely. You are moved to the new quota system, but you keep your current price.
Your per-seat price is grandfathered in at $30/mo per Developer seat indefinitely.
Every existing paid subscriber gets a free extra week added to their current plan. This means your next renewal date was extended by 7 days.Use that week to try the new quota system and see how it maps to your actual workflow.
Your annual subscription renewal date will be extended by 7 days for the trial week. If you decide to cancel, you can request a refund for all remaining months on your subscription.
Enterprise Self-Serve customers continue under their existing billing agreements. These changes do not affect you at this point.
Enterprise customers continue under their existing billing agreements. Reach out to your account team with any questions.

​
Add-on credits & extra usage conversion

Quotas replace the built-in prompt credits that were part of the previous credit-based system.
All add-on credits are converted into a dollar amount of extra usage at the rate you paid for them. Since prompt credits were sold for 0.04/credit,forevery250add−oncreditsyouhadremainingonyouraccountyoureceived0.04/credit, for every 250 add-on credits you had remaining on your account you received 10 in extra usage balance.
Yes. After conversion, you can request a refund of your unspent extra usage balance at any time through support. You’ll receive the equivalent dollar amount back.
No. Credits are converted at exactly the rate you paid. You can either use that balance as extra usage going forward, or request a refund.

​
Other questions

A successful referral (when the referred user pays and starts using their subscription) grants you $10 worth of usage.
The student discount now offers the Pro plan for $10/mo.
If you previously purchased SSO as an add-on, you keep SSO access under your grandfathered plan. New Teams plans do not include SSO — it is now an Enterprise-only feature.

​
Token pricing example

To show how token pricing works in practice, let us walk through an example conversation with Cascade using Claude Opus 4.6:
RoleMessageTokensNote
UserRefactor @my_function20kInput (cache write). Note: Incl. full shared timeline, editor context & system prompt.
WindsurfLet me first analyze my_function to come up with a plan to refactor it.1kOutput tokens.
tool_callAnalyze my_function23kInput (cache read) + Input (cache write).
WindsurfHere is a plan to refactor my_function […] do you want me to continue with implementing?2kOutput tokens.
UserYes, continue.46kInput (cache read) + Input (cache write).
tool_callEdit foo.py50kInput (cache read) + Output tokens.
tool_callAdd bar.py56kInput (cache read) + Output tokens.
WindsurfI am done refactoring my_function. Here is a summary of my changes: […]2kOutput tokens.
Total200k
The actual per-token cost can be calculated based on the model pricing table page.