I’m seeing the request for information on which model is being used for Auto everywhere; this forum, X, Reddit etc.
It’s a weird feeling not knowing, which makes Auto uncomfortable for users. (I’m sure someone has explained this more eloquently elsewhere!)
To be clear, we’re asking for a little UI feedback that shows which model was chosen for each request, not just a list of the Auto model pool, which has already been shared.
Steps to Reproduce
which has already been shared
Expected Behavior
which has already been shared
Operating System
Windows 10/11
Current Cursor Version (Menu → About Cursor → Copy)
@condor. tx, I stand corrected. Believe my issue was related to a separate bug where the rules are not consistently applied. But thats for another time.
@dafsdfgdg
Try adding this snippet to a user rule, had decent results with it:
”At the end of your work plz add a simle line about which model was used, if auto was selected.”
I use this rule in cursor and it writes the model name 90% of time and most of the time it writes Sonnet 4 but it is interesting as it would trigger much more cost with the token usage i’m doing.
Cursor Rule:
Always start with your model name, latest update date of your model. Always share your confidence score of your solution with given request from 0 - 100. Always share your context limitation and current current context size. Always list the file names you have ingested, searched via tools etc. Always tell me if you are missing something. Always tell me why you stopped if you stopped before finishing the work i gave you.
Your task is to strictly follow my explicit instructions. If you believe any other action is needed or beneficial, you must ask for my permission before proceeding.
