For the price we are paying, the new requests limit is veryy low didn’t even finish 6 days of usage ( with the same workload i used to get at least 2 weeks or 3 weeks of the month to hit the limit. ).
hi @MASMOUDI_MOHAMED and thank you for the post.
Cursor team is looking into details, could you share which models you are using and if possible share a Request ID with privacy disabled as this would allow checking what happens within the request?
claude sonnet 4 ( same as the last 2 months i think ). i can’t find a request ID, but the point is the new limit is veryyyy low and not worth the 20$ anymore.
Imagine this: I hit the limit of my ULTRA plan after 5 days using it. No agents involved. Nothing unusual.
Cursor has become a total rip-off.
I was reminded 4 days after I paid, and again today (5th day). I don’t know what the point of a $20 subscription is. I didn’t make 225 requests, not even 100, and the bill is already $19.X. request id:d109434d-ffdf-4525-92b9-77f4bed9a03f
you have to switch to old pricing
this much better
switched to Claude code instead. much better
The Pro plan is totally useless. My billing cycle started on the 17th, and after just a few hours of usage, I was already hit with a rate limit. The Auto mode is unlimited, but the quality is extremely poor. If the new billing is token-based, then at the very least, they should provide transparency on token usage, how a couple of prompts can drain a million tokens with o3 or Claude Sonnet 4, while Auto consumes significantly fewer tokens compared to other models, even though it generates a large amount of text. If the new plan is based on token usage, they need to revise the plans and clearly inform subscribers of what they are getting. I mistakenly switched to the new plan in the middle of the month, and now I cannot go back to the old one.
1000%. I use claude code and havent looked back. I plan on keeping cursor or the other AI product to allow for other models when needed at limited capacity and use claude code for everything else.