In your blog you stated that existing users can choose to stay with the 500 request limit method. How do we choose this option? Contrary to what you’ve written, the new pricing model is far less “generous”. Models like Claude 4, which consume 100 times more cache tokens than input/output tokens, have become practically unusable under the new system. Additionally, the previous request-based model made it easy to track usage, by watching how many requests were charged per model and view request history. The new system, based on blurry rules, has made this nearly impossible.
I can not choose it. With the new model I hit limits WAY TOO FAST. If I could atleast use the slow mode.. But nope.. I don’t know what to do right now my wallet is not so big
What also bothers me, is that i can not really see how much I can use. I see used tokens, ok, but how much do I have left? It was so good before, I could see ok I have 200 fast requests left, now everything is very confusing. I’m sad, NGL
Hey, firstly I want to apologise if things have not been clear for you all here.
While many users are now able to get >500 requests each month, using the top models will always consume your usage quicker than cheaper models. For example, with the new system, you can now get >500 Gemini 2.5 requests each month, as well as unlimited use of the ‘Auto’ mode selector.
The goal here was twofold: to simplify and to improve the fairness of our pricing.
On simplicity, the system is now much less “involved” (for lack of a better word). Our hope is that users will settle into a tier and a usage pattern that fits it. For those who are heavily rate-limited, it likely means they were relying heavily on usage-based pricing on the old plan, and would likely get more benefit out of upgrading.
For fairness, your consumption of your rate limits now directly matches your usage of the model - bigger conversations and better models = higher usage of your limits. Many $20/m users will now be able to use Cursor more, with unlimited Auto requests, allowing more usage for our most cost-sensitive users who have less lock-in to the best models.
We are still working on improving visibility here, but we already have some in-app alerting to show you when you are approaching your rate limits!
