Questions about on-demand / usage-based pricing

Support Help
1

Hi everyone,

I have a few questions about how the on-demand (usage-based) pricing works in Cursor once the included plan limits are exhausted:

  1. How much does on-demand usage cost?

  2. How exactly is it measured (e.g., by token, by request, by model)?

  3. What are the specific prices or rates for the different frontier models (like Claude 3.5 Sonnet, GPT-4o, etc.)?

  4. Why isn’t this explicitly clear on the pricing page? I’ve found it a bit confusing to figure out exactly what to expect.

  5. Where can I find more detailed information or official documentation about this?

Thanks in advance for your help!

5

Hi Fernando,

On-demand usage is billed at API rates listed here: Models & Pricing | Cursor Docs

The best way to track your usage and spend is here on the Cursor Dashboard: https://cursor.com/dashboard/usage?from=2026-05-01&to=2026-05-20

By token usage and model choice, see Models & Pricing | Cursor Docs for details.

This is also available here: Models & Pricing | Cursor Docs You can click on a model name to show more info, for example: Claude 4.7 Opus | Cursor Docs and you can see pricing details, context size, and more.

If you’re worried about spending too much money on on-demand usage, the best thing is to set a reasonable On-Demand Usage limit here: https://cursor.com/dashboard/spending and then track your usage. For example, if you set an on-demand spend limit of $15, you won’t be charged more than $15 above your regular plan (subscription) costs.