Gemini pro - I'm having a hard time fulfilling your request. Can I help you with something else instead?
I am a Gemini pro subscriber and have been experiencing constant, disruptive generation errors starting today when submitting simple prompts, specifically around parsing network IP data, conducting research, and structuring the output into a table. Even when opening a brand new chat session and submitting a prompt on the very first turn, the system repeatedly fails and outputs generic refusals such as stating it is having a hard time fulfilling the request or that it seems to be encountering an error. Even saying just "Hello" is enough to receive those errors.
This issue is occurring across both Flash and Flash Lite models, and I have not attempted Pro as it is unnecessary overkill for such a lightweight query. My quota is nowhere near exhausted, with daily usage sitting at roughly two percent and weekly usage at three percent. I have already performed extensive local troubleshooting, including hard cache refreshes and switching over to an alternative network to ensure local firewall or DNS filtering rules are not interfering with web socket streaming.
The only unreliable workaround (usage via chrome on windows 11) at present is making arbitrary edits to the text input, such as adding a trailing space or deleting a full stop, and resubmitting until the model randomly decides to process the request instead of erroring out. Having to repeatedly modify and resend prompts makes the service completely unusable for basic tasks. I would like to know if anyone else is experiencing this issue, and whether there is an active backend service degradation or a newly introduced safety filter false positive affecting the Flash endpoints today?