Hello,
I switched to the Pro subscription. I’m happy, but there are a few things I don’t understand. Although I looked in the forums, I still didn’t understand much.
1- What happens when the limit of 500 requests per month ends? Can I send unlimited slow requests?
2- When I downloaded the application, in the settings, in the model selection section, o1mini - o1-preview was automatically selected. What should I do to use this o1-mini for free? When I bought a pro subscription, it said 10 times for free.
3- I have never tampered with the model selection side, I did not select and remove it. Do we set paid free use or fast slow use from here.
4- The basis of the questions I asked above is that every time I write a code in the application or ask something in chat, 1 premium request goes. Is there someone who can explain a little more detailed?
Obviously, is there only a difference in the number of requests between the normal version and the pro version?
Visuals are attached. Good work everyone 
Finally, when we choose from here, no matter which one we choose, is 1 premium request still gone? o1mini limit 10, for example, is there a warning that this is full? I am asking to avoid being billed extra.
Hey there!
Based on my experience, here’s how Cursor’s pricing and model usage work:
-
Usage-Based Pricing: If you don’t have this enabled, Cursor won’t bill you extra for any requests. Worst case, your request may fail if you try to use a specific model.
-
Model Pools: It’s unclear which models fall into which pool (Pro or Free). I’ve used o3-mini-high multiple times without seeing an increase in either meter. Some models like Deepseek v3 might be part of the unlimited pool due to their API costs.
Now, let me address your specific questions:
-
Exceeding 500 Fast Requests: If you reach this limit, your requests will take longer, depending on current demand and the number of requests you’ve made beyond 500. Essentially, you’ll be placed in a slower queue.
-
Using Models Without Extra Billing: With usage-based pricing off, try using any model. Cursor won’t bill you extra; worst case, you won’t be able to use it. For example, o1-mini should work just fine.
-
Selecting Models: Feel free to select any model you want to try. Some may not work, and it’s unclear which models count towards which pool. I often use a model and then check my Cursor profile to see which pool’s usage has increased.
-
Small Free Models: Try using small free models like gpt-4o-mini, which is quite capable for tasks like inline edits with Ctrl + K.
Regarding your question about Pro vs Free versions, if you’re referring to premium models vs normal models, the distinction is mostly related to their API pricing. More expensive and capable models have a limit of 500 fast requests; beyond that, you’ll be placed in a slower queue.
Hope this helps!
Followup: The models and what their rates are can be seen here:
1 Like
No, 4o-mini and cursor-fast e.g. are part of the unlimited models
1 Like
