Anybody else notice the latest Gemini model is just really bad. Forgets code, removes instead of fixes, gets stuck in loops, unable to complete a task..
Not just in Cursor but in Gemini Pro too.
Almost as if it lost 50% of its intelligence.
The model was updated in early June, quite a while ago. Personally, I haven’t liked the hidden thinking process since then; before, it was very clear how intelligent it was.
1 Like
The June 5 version will be deprecated soon.
1 Like