cursor new pricing model is good but i didn’t find the reset time
1/ Does cursor have a specific time for model reset, cursor is a tool that many people use for their daily programming work but I have never found the specific time
If not, could there be a separate dashboard page that shows the reset time for each model rate, so that users can plan their work time allocation?
Currently at least 10 different threads are asking in different way to get any way of knowing how Cursor is calculating any limits, and Cursor folks are adding the button to switch from Pro to Ultra in an easier way. That is sum up of all discussions on that topic since new pricing model started.
So no, that is hidden better then state secrets are.
Yeah, I got also no details on that.
From my personal Pro usage, I ran today Claude 4 Sonnet Thinking Max for 3.5h non stop which made about 560+ events in Cursor com > Usage list (all included in Pro, with charge shown as -). I had Usage Based Pricing disabled to see how long it would last me and finished the very detailed task with several thousands lines of changes in that time with minimal supervision. Havent yet reached the limit.
With Opus etc it would sure reach easier but Sonnet is running well.
Note that I manage context very well, minimal task details (just required info, no rules, no attached files, it fetches all it needs when it does)
If you @qwert @nightcoder dont mind sharing how was your usage until you reached rate limit?
Which models, mode, duration…
I was on 113 request when the planning changed. I used it with Sonnet 4 after it intensively for maybe 3-4 hours and reached the first limit. Enabled usage based pricing just to see what’s gonna happen, and spent 2$ with 3 prompts in the first hour. I did see that at some point it is using that money, at another it is not, but couldn’t narrow down anything.
My code is written by myself, structured by myself, simple Cursor file on top of it, DDD to the bone, so Cursor is able to do the tasks for myself. For some reason it is/was halucinating more then I am used to (I am really reviewing everything), asking me if I want it to do what I told it to do, but hey, another request to pay…
So today I am using my work account. It still shows 500 requests, old dashboard, started with 0 requests this morning. Used it maybe in total for ~5hrs today, just checked, it said I used 213 requests today over 70 chats. And now I am scratching my head thinking what 70 chats really means… 70 prompts? At least 20 times saying “you are halucinating”? This rubs me really the wrong way…
Will try tomorrow more intensive coding. The ‘events’ are more than just requests but any event that consumes resources, but its the closest I see to old usage logs.
While I was typing previous post, I was waiting for it to finish something. Just checked - I am on 219/500 requests, still on 70 chats.
hi Thanks for the reply Yes the new version is very good Being able to use max mode with confidence, in the last version I consumed a hundred fast requests with only one O3. And the current new version I used it many times and still haven’t reached the limit
It would be nice to know exactly when the limit resets, to be able to utilize max mode better.
I just using the new pro plan today (first time since the change) and I ran sonnet thinking max for surprising periods of time waiting to see how long it took to hit a wall. I never got a rate limited message but occasionally tokens/per second slowed to a crawl. I dont know if that’s related to usage or just typical latency spikes. I wonder how actively they are tuning the rate limiting and experiementing with throttling tokens per second once you have saturated usage over their limits