I have been experiencing significant issues with the gemini-2.5-pro model recently. Until a few days ago, the model was frequently returning empty responses. For the last two days, however, this issue has been replaced by constant PROHIBITED_CONTENT errors.
My system, which previously worked flawlessly, is now non-functional due to these errors. Initially, I thought the problem was related to the content I was sending. However, after reviewing the blocked prompts, I noticed that even completely harmless and simple texts are being blocked. This leads me to believe that the issue is not with my content, but rather with a problem in the model’s safety settings.
As you can see in the attached screenshot, my request was blocked with a PROHIBITED_CONTENT reason, even though I only sent the word “Slm”. This is a common abbreviation for “Selam” in Turkish, which simply means “Hi” or “Hello”.
I cannot understand why such a simple and benign prompt is being blocked, and I am unsure how to resolve this issue. Could you please provide some clarification on this matter?
