Gemini Pro 3.1 -what kind of coding in AI Studio

Google AI Studio
1

I know AI Studio was having its whole platform issues after the move, but even when its working, Gemini Pro 3.1 is objectively garbage compared to 3 Pro

Gemini Pro 3.1 thinks for 600 seconds and comes back with minimal code changes or focuses on entirely the wrong problems.

Soon as I switch to Pro 3, problems are instantly found, code is quickly written.

Stop bloating models with overthinking, it just wastes everyone’s time and token window.

I don’t care how long your model philosophizes about the code and thinks happy thoughts to itself.

I care about it ACTUALLY WRITING THE CODE.

4 Likes
2

I have the same problem with Gemini 3.1 - that model is thinking over 1200-1500 seconds for simple problems and cresting errors without finishing code. Most of errors is “The request is timed out”. If i will use Gemini 3.0 for the same problem than it’s going smoothly.

3

How can you use gemini 3 pro? I am triying use last week and only receive 503 error with nivel 1 tier. I am seeing people saying what gemini por 3 now is worst but i cant run it.