Gemini 2.5 Flash quality degradation

I may be going crazy, but I noticed a degradation in the overall quality of Gemini 2.5 Flash. In addition to the price increase, I use it more without reasoning, which I thought was cool to have separate prices for reasoning and without reasoning, which we no longer have.

Now Gemini 2.5 Flash has become expensive for the project I was working on. In some tests, I noticed that it looped several times, something it didn’t do in the preview.

Did Google change the quantization of the model? Any chance of returning to previous pricing?

1 Like

Hi @Gabriel_Scatolin,

Welcome to the Forum!

Thank you for bringing this to our attention. We truly appreciate you flagging this issue, we will file a bug internally. To help us with prioritization of this issue, could you please provide more specific details about the quality decline including examples of tasks or outputs where you’ve noticed a difference?

While I don’t have immediate information on quantization changes and pricing, I will inquire with the internal team.