Plan: Cursor Individual Pro ($20/mo) Billing cycle: Apr 16 – May 16, 2026
Hi team, I wanted to share what I’m seeing on my end and get some clarity before drawing any conclusions.
Auto mode is currently showing “Total usage limit reached” even for simple tasks like reading a file. I’ve pulled up my dashboard and both pools are at 100%:
API pool: 140.5M tokens (100%) — this was exhausted through manual model use (claude-opus-4-7-thinking-high, gpt-5.3-codex). These were intentional selections on my part, so I understand that pool being used up. And this was exhausted quite a while ago, around 1 week before.
Auto + Composer pool: 288.9M tokens (100%) — Auto alone accounts for 235.1M of that. And the main 100% usage was exhausted 1 week ago as well, so it’s not something which reached limit today, but rather was reached at the threshhold before.
On-demand spend: $0 — no spend limit has been configured.
What I’m trying to understand is the Auto + Composer pool exhaustion specifically. I may have missed something in the documentation, so I have a few questions I’d appreciate some help with:
What is the actual token ceiling for Auto mode on the Pro plan? I want to make sure I have the right baseline before drawing any conclusions — if this is documented somewhere I may have missed, happy to be pointed there.
When the Auto pool is exhausted, is there a reason it blocks entirely rather than routing to lighter internal models for low-cost tasks like file reads? I’m trying to understand if this is expected behaviour or something worth the team looking into.
Is setting an on-demand spend limit the recommended path to restore access before the May 16 reset, or is there another option I should be aware of first?
Happy to share any additional usage data from my dashboard if it helps diagnose this further. Thanks in advance.
Hi @Techerudite_RNTL Thanks for the post! I think I probably answer a few of your questions here:
The immediate unblocker would be for you to enable on-demand spending with a limit you’re comfortable with. You can then continue to use Auto, Composer 2, and any named model that will draw on your on-demand usage until your cycle resets. Auto and Composer 2 are going be the most efficient and cost-effective options to use if you want to contain on-demand spend.
I udnerstand your insight, however, both the auto + composer as well as premium quota was exhausted for me over a week ago. It was both 100% for over 10 days, and the ‘auto mode’ was working for me as expected, just with ‘slower responses’ which i think may be soft rate limit instead of a complete blocker.
It was just yesterday that i reached this hard limit where i was not allowed to perform any chat / auto mode operations. So i am just trying to understand exactly how the limits are placed, and how I can review these from my side. Also i was under impression that ‘auto’ mode was ‘unlimited’ so just want to have clear understanding of how the quotas and limits work.
For example, i have got another account where 100% quota has been exhausted, and the same thing happened with that one as well. Yesterday i was able to use it correctly, and suddenly today it has stopped working. So i am just wordering if it is something to do with limits, the no of days remaining for the billing cycle or something else entirely.
Auto mode is not unlimited but bigger plans have much more usage. So Pro + has more auto usage than Pro, Ultra has more auto usage than pro + etc. It’s very generous! I recommend upgrading to a higher usage plan.
I understand regarding having more usage. I am just looking to have a tangible metrics or figures that will allow me to understand what is the ‘usage limit’ so it can be managed properly without getting stuck on this hard block. As the usage was shown as 100% on the cursor dashboard before as well and i was still able to continue using auto mode. So if there is a way to understand how and when these thresholds are going to be hit, then it will be helpful to keep an eye out and plan accordingly.
If you can shed some light on this if you are aware, then that will be appreciated.
@Techerudite_RNTL Good question, and I think I can provide some clarity here.
There are two pools that reset each billing cycle: an Auto + Composer pool, and an API pool (for named models). When your Auto + Composer pool is used up but you still have API pool left, Auto and Composer keep working — that usage just draws from your API pool. Once both pools hit 100% and on-demand usage is off, you’re blocked, which is the point you’ve reached now. You can always enable on-demand usage and set a modest limit to keep working.
On a tangible number: we don’t publish a fixed token ceiling for the Auto + Composer pool, so your dashboard is the source of truth. The most useful thing to track is your burn rate — at cursor.com/dashboard → Usage.
To stay ahead of the hard block: enable on-demand usage with a spend limit you set (a controlled cap, no surprises), or upgrade to Pro+/Ultra for a larger included pool. To stretch what you have, prefer Auto/Composer over premium named models.
