What are the quotas on non-premium models?

Hi all,

on /pricing I see my quotas for GPT-4, GPT-4o, and Claude 3.5 Sonnet/Haiky, the “premium models,” and I see that calls to 4o-mini or cursor-small are unlimited. But I would like a clearer idea of the monthly quotas for other models.

I read somewhere that the calls to o1 and o1-mini are limited to 10/day (is it that right?) but I don’t see this documented anywhere.

Are there other models with a daily quota? And are all other models (non-premium, not listed previously) usable with no quota?

Hey, you get 10 free requests to the o1-mini model. The o1 model costs 40 cents per request. I’ve attached a screenshot with the prices for all existing models based on usage-based pricing. Note that you can now also purchase premium models with this system, and you don’t need to buy in batches of 500 requests. If you exceed your quota, you can enable usage-based pricing for premium models for your convenience.

Thanks, I see where it is now in the pricing page. A question about the usage-based pricing: if I understand correctly, it’s 4 cents/request for the normal premium models, but only if I go past my 500 monthly quota?

I don’t understand here where it says it’s billed every 500 requests. Does that mean that if I use 200 extra requests one month and 300 extra the next month, I get charged nothing at the end of month 1 and $20 for month 2?

Can someone post a clear link where that info is shown on the pricing page because its not there.

on the pricing page you have to go click on ‘enable usage-based pricing’ and click to expand the pricing details

Ok thats not acceptable. Pricing must be visible clearly on the public pricing page. @danperks
Not clearly showing prices can be a huge issue in many jurisdictions across the world. People need to know the price before they commit to something and not when they enable that feature.

Yes, if you exceed 500 requests, you can enable the usage-based payment system. You are only charged for the requests that exceed your quota. When your quota is exceeded, you’ll be switched to slow requests. If you need fast requests, you can increase them in the settings and pay $20. If you don’t need 500 requests, you can enable usage-based payment and use as many as you need. Keep in mind that unused requests expire at the end of each month, and charges for additional requests are calculated separately for each month.

hard disagree on the $20, so many posts show what issues happen when users on certain plans do that addon. Its not about the price. its about how cursor internally handles that and what issues cursor has internally.

The pricing of the Pro plan is clear on the website.

Usage-based pricing is an add-on, on top of the Pro plan, and requires a Pro plan to use, to we show it only to uses who have Pro already to avoid confusion. We may, in the near future, add detail pricing to the documentation.