Gemini keeps forcing 1.5pro

I’m doing some light coding with Gemini in Google AI Studio. I have a Python script that uploads a document and sends it to the model using my API key. The problem: Gemini keeps rewriting or correcting my code to use gemini-1.5-pro, even though I explicitly specify Gemini-3-pro-preview.
Whenever I ask Gemini for help, it replies with code using 1.5-pro as if it’s the only valid model. It even “helpfully corrects” me and removes the 3-pro reference.

Anyone know why Gemini ignores the requested model and insists on downgrading it?
Is Gemini-3-pro-preview not officially supported yet in Python, or is this just the assistant hallucinating?

Any insight appreciated.