The problem is that you completely stopped before even reaching 225 times per month. If you don’t believe me, check my account
lol but danperks is not really like following someone rules? or he is the leader?
but even with the claude sonnet api key is too, You have to pay a lot for that we need to stop just reading or talking s/ just everyone make alot of post mass post with your accounts legal without breaking anything and if they dont do anything if they with 100 of real post about the limited unlimited in that case yall can mass spam with bots but first make them know we are here legal always legal and everyone need to Raise your voice if they, after stealing money, after changing everything, don’t want to do anything and ignore 100 posts with the same complaint, then it would even be legal to make mass spam posts. think about it @0xHACKS
LOL! I paid $60, and I’m still waiting for next month to be able to use it. Anyone else believe and will continue to pay for the ultra plan?
Your usage limits will reset when your monthly cycle ends on 8/3/2025.
4 Likes
You’re right that Claude Sonnet’s API isn’t cheap, $3 per million input tokens, $15 per million output, but it’s still cheaper than paying Cursor’s hidden 20% markup if you don’t bring your own key. The real issue is lack of transparency.
Raising voices legally and loudly is fair, mass posts with facts work better than emotional spam. Cursor won’t act unless the pressure is public and consistent.
3 Likes
Nice bootlick, but nobody’s above criticism, not even a dev. We’re talking facts, not fan clubs.
1 Like
The cursor unicorns trend appears to be over.Cancelled my monthly subscription.
4 Likes
Please, I need clarification on Pro subscription is it base on usage of token or it is monthly usage and the hrcursor should please reply to my mail.
You deeply offended him and hurt his faith in an honest and transparent pricing plan, which he sees as a generous offer.
He’s deleting comments and threads to keep people from straying off the righteous path.
I’m sorry, but you’re using fake math. Simple queries for working with a single file can be displayed as 400,000 tokens, but this is an INCREDIBLE amount (there are 4 notes in memory, 1 file per 100 lines is attached). Sometimes Claude uses 1,200,000 tokens for a single request, which is just an incredible amount to work with the model directly, for example (the context window is only 200,000 and you can solve the task 3 times there). So, the question is not whether you “used $500” to reach the limit, but that we have no idea where these tokens come from! After all, just to compare my work, I can work 3-4 times less time than I did just a month ago. I haven’t changed my work style or project, but on the 15th I reached the limit. Previously, if this happened a few days before the end of the period, there were always enough “slow” requests. Last month, I received a warning that I had reached the 90% limit and used the remaining 10% for another week by switching to o3. This time, I immediately used 100% of the limit twice as fast. This does not correspond in any way to the experience of the past 4 months of working with Cursor, so “this is better for most users” is another unsubscription without facts (the fact that 70% of users do not complain on the forum is not an indicator of what is better for them)
2 Likes
Last month, my Cursor generated nearly 300,000 lines of code. This month, after generating just 40,000 lines, I was notified that I could no longer use it. And this is only half a month into my reset cycle.
I have to say, the Cursor team did it. This is a huge success. 
As a yearly pro user, I am deeply frustrated with my choice. I will not be considering Cursor again next year. It’s just a middleman that arbitrarily changes its pricing model and dumbs down the AI’s intelligence.
5 Likes
This is the response I received from the Cursor team. They introduced a usage quota, but initially emphasized that it would be unlimited. And when customers complained, they said most users wouldn’t hit the limit anyway. It’s ridiculous coming from a company valued in the billions
1 Like
it’s a big issue at the moment
1 Like
same here man
this is a bug, definitly, I literally did one reqquest to center a div and you can see now that suddenly I’m using MILLIONS of tokens
I can’t spend at that rate is a complete scam, literally 3 mins of ussage 3€ and is a normal thinking model not max or anything! complete scam
8 Likes
@danperks please provide some explaination
Funny how it suddenly decided to start working harder all of a sudden.
@danperks I am one of those users that doesn’t seem to be affected by this as much. But even I, just learning C# and web-development programmming in very, very small projects, am seeing pretty stunning levels of context usage without being able to pinpoint why. I can see how much has been used for input, output and the cache read. And “Cache Read” appears the clear thief here. I can’t help but think that something is going wrong in how agents are processing information.
This IDE works for me, but I am wholly unhappy with the communication that has been transpiring with regards to the community feedback on rate limits and token usage. We eventually got feedback on the rate limits, actual useful such. But it took longer than it should have, and this thread is quite simply making me uncomfortable here.
As another one who has worked in customer support, I recognize that you’re in a tough position. But if I had been in your shoes this would be a moment to either…
- …have a very clear discussion with leadership on being transparent with he community.
- …stop gaslighting customers by only responding as if they hadn’t brought up perfectly valid points regarding massive increase in token usage.
- …possibly quit your job, for your own sake. This communication cannot be fun to do for the average person.
If the company can’t actually handle this community’s frustrations I will be moving to another platform in a few days. My subscription runs out in a little less than a week, there is competition out there that’s definitely worth using instead.
7 Likes
Yes indeed is so confusing I am using the ultra and now said I have reach the limits too