And it was smooth sailing for a month. I have mostly used Auto, occasionally - Sonnet or Opus.
And I’ve used about 90% of my monthly quota between Feb 21st and Mar 21st. I’m on a Pro+ yearly billing plan.
I’ve enabled the setting that shows usage as “You’ve used N% of your usage limit” under the typing area in the Agents tab.
During the 1st month I’ve noticed that using Auto did not affect that percentage at all. I was under the impression that Auto is free and I was only charged for using more advanced models like Sonnet.
As the 2nd month of usage started for me on March 21st, I have noticed that:
- Using Auto is no longer free. I’ve mostly used Auto for 3 days and ended up at “You’ve used 6% of your usage limit” in the Cursor app.
- On the Dashboard in the Spending tab I still see “1% Auto and 1% API used”, which would be in line with my past month usage for similar workload. Why am I seeing 2% here, 6% in the Cursor app? Is one of these displays bugged or are they showing something different?
- It seems like the model used for Auto has changed. It seems to be generating text much faster while it’s thinking.
I was semi-satisfied with the previous Auto model. It was doing okay with good instructions given.
Can you provide any insights on why the usage in the Cursor app on the “You’ve used N% of your usage limit“ is growing so fast compared to last month (while using Auto)? Was it some “1st month grace period“?
Discovered that I can install a browser extension to monitor Cursor usage. Here’s what the extension says:
Your plan refreshes in 29 days
Current month
42 requests
You’ve made 42 requests during the current billing period.
Usage
1% used
Tokens: 13.5M
Cost: $5.98
Usage Breakdown
Auto + Composer
1.2% used
API
0.8% used
Models Cost Breakdown
default
5.9M tokens
$2.82
composer-1.5
6.8M tokens
$2.23
claude-4.6-sonnet-medium-thinking
777.0K tokens
$0.92
composer-2
20.4K tokens
$0.01
Usage-based pricing:
Disabled
In the meantime, in the app I am at “You’ve used 8% of your usage limit”.
It counts it as $5.98/$70
But what’s the meaning of Usage 1% used in the Extension then?
And “1% Auto and 1% API used” in Dashboard/Spending?
What would happen when I hit “You’ve used 100% of your usage limit” usage in the app?
And another observation - during the 1st month, I was checking both Usage in the dashboard and usage in the app. What was interesting back then - usage in the dashboard was showing higher numbers than the usage in the app. I remember that at some point my Usage in the dashboard was $25 and my usage in the app was $12. I tried to figure out which of these numbers mattered most and back then I found some resources stating that usage in the app is more important and usage in the Dashboard might be higher, but it is irrelevant for the actual usage limit, which should be checked in the app. Back then Dashboard included cost of auto and other models like Sonnet, Opus. App seemed to only be affected by custom models (non-auto).
This month though, my usage $ amount in the app is exactly equal to my usage $ amount in the Dashboard. Any changes related to this lately?
Hey, thanks for the detailed report and for updating the thread with the outcome.
The mismatch between the in-app usage % and the dashboard is a known sync delay issue. Sometimes the in-app counter temporarily shows higher numbers, then corrects itself later, which matches what you saw.
A couple of clarifications on your questions:
The dashboard (Spending tab) is the most accurate source for usage. The in-app counter can temporarily lag behind or jump ahead.
Auto mode has always used quota, but depending on which model Auto picks, the cost can feel different.
When you hit 100% of your usage limit, you’ll get a warning and models will switch to lighter ones, or you’ll need usage-based pricing.
The team is aware of this desync issue. If you see the mismatch again, use the dashboard as the source of truth.