“Agent Review” feature seems to use premium “non-auto” LLMs? Can I set it auto?
I am out of “premium” quota so agent reviews fail.
Failed to run review: insufficient funds (request ID: efa099e3-4b5e-4314-84a3-6bf373fbaa79)
hey @Satoshi, you can modify the model used by agent review in the IDE settings, choosing quick really lower the end bill.. tho I’m not aware if it use premium quota or not, you can also check Agent Review | Cursor Docs
hi @Tom_Coustols thanks! yes, I already set it up that way and it seems to use premium quota as I ran out of premium quota while still having “auto”. So, I assume it uses premium… I was not able to find any info how to switch it to “auto”.
well 
, it seems that as of now, switching agent review to auto isn’t actually possible.. tho something I did a while back was to work with the headless cli that cursor allow us to use, and I put it in my github action and it was actually kinda running like what the agent review does, all that using auto!
thanks! I guess we also do not know what LLMs they use for agent review or the system promp? I found it quite good, especially also that it actually checks if I have something within /rules. Whatever they did to make it look at /rules consistently they should also apply that to the normal “agent” mode. Because there it keeps “ignoring” the rules 
but I am aware that this is a known issue already well discussed in the forum.