Usage shows wrong models, is this a bug?

[haris-musa]

I have “claude-3.5-sonnet” selected all the time but my usage says that I have used 2067 fast requests of gpt-4o mini / cursor small. What’s going on here?

[deanrie]

Hey, here’s why this happens.

[SOLVED] Unintended Use of GPT-4o-Mini

these gpt 4o-mini / cursor-small requests were being incremented when our apply model was used (for legacy usage accounting reasons). the main chat/composer response was still from whichever model you selected in the app just pushed a fix to stop counting apply usage here, let me know if you still see this!