What happened to the AI of the Cursor models?
I use Claude sonnet 3.5 and Deepseek r1 all the time, it worked fine, but for a few days now there are serious problems. The models do not perceive context well and very often hallucinate mixing their fantasies with the existing code base, their answers have become extremely stupid. What is the reason for this degradation? In their original chats, such problems are minimal.
can you try re-indexing your codebase once? i faced the same problem a few days ago and re-indexing the codebase and starting a new chat made the responses better
Thanks for the tip! Could you give me a hint on how you did it?
go to cursor settings (the settings icon on top right corner of window)
then go to the features tab, scroll down and you’ll find a Codebase Indexing column
Just click on the “Resync Index” button there and you’re good to go!
I’ll give it a try, thanks again!