When I hit one GPT-4 request on the cursor.sh, It count more than 10 on my profile

When I make a single GPT-4 request using the cursor, and then review my profile’s GPT-4 usage, it shows an increase of over 10. For instance, prior to the request, the count was 53, and post-request, it rose to 66.

That is not good! Went ahead and sent you a refund for this month. Certainly want to fix whatever is going wrong here.

Does this happen only when you use a particular feature?

This issue persists whenever I engage with GPT-4, whether through shortcuts like ⌘+K or ⌘+L. In some instances, it count just one usage, but more often than not, it counts multiple instances, occasionally exceeding ten.

Do you still see this happening now? Just pushed a change on our end.


Thank you so much, its working fine now.

1 Like

Today again facing this issue.
It’s again count 12 on my profile by calling just one request on cursor.sh

I’m not sure if it’s related, but today, the queues (after the quota is exhausted) seem to be larger and slower than usual.

Is this happening on cmd-k requests still? Or was this from another feature (maybe /edit)?

Just made a change on our end - let me know if you see this again!

It occurs when I send a prompt in the cmd+L window, receive a response, and then proceed to click on the ‘apply diff’ button.

1 Like

Alright, I’ll inform you when I will encounter this issue.

1 Like

Enabling “Reasoning step” in the advanced codebase context settings also uses multiple GPT-4 requests.