only finish a few todo items in the todo list, then stop there.
followed with a icon with token used.
this makes me frustrating, waste my requests quota.
Steps to Reproduce
request in chat with gemini-2.5 pro model
Operating System
MacOS
Current Cursor Version (Menu → About Cursor → Copy)
same thing happens to me, happens on GPT also, it does 1 or 2 things then it says it’s going to continue but then it doesn’t.. the model even says i’ll continue fixing {whatever} and then it just stops. while it’s not a big deal it does waste time and of course it wastes money by having to constantly tell it to continue which consumes another request
GPT5 doing same thing even worse it says it’s going to do something but then doesn’t do it. Or it says it did it and didn’t. Also even if it does do something it changes 2 lines in a file , says it’s continuing to change more but then stops. By stops i mean i get the chatbox back , the request ended, which means i have to prod it and waste many tens of requests to get it to do something.