Wait, are you kidding me? After the update i now have to wait 3 days? Why? What’s wrong? What i’m paying for? I burn token/usage because of errors and i’m rate limited?
How can i get in touch with google support?
Nah every single Claude model, and Gemini 3 Pro now behaves exactly the same as Gemini 3 Flash - 100s of tokens in incoherent rambling of a “reasoning” and then no actual work or worthless output
”I am now meticulously wasting your tokens on fake reasoning while providing no actual work…”
Edit: Well it seems to work better now, guess it was being routed to Flash even if you selected Pro. But still, capability wise it is no Claude (especially Opus).
This is so infuriating. 90% of inferences on any claude model end in this non descriptive error.
If you are incapable of serving these models then remove them and don’t mis-advertise your product as having them. Cause they DO NOT WORK. For literal weeks now, not an occasional outage.
Not that that will be particularly helpful, since you often route Gemini 3 Pro to Gemini 3 Flash instead.
Having this exact same issue all day. I am an ultra subscriber and there is zero support. This product is terrible. Why is there no support or indication if the problem is on our side or theirs?