I am reporting a critical stability issue with the Gemini 2.5 Flash API that is causing significant downtime for our production application.
We are consistently receiving 503 “Service Unavailable” errors with the message: *“The model is currently experiencing high demand.”
*
Request: Could you please investigate the capacity health for the Gemini 2.5 Flash clusters associated and **Gemini 3 Flash preview
Please tell any solution as well**
yes! facing the same issue today. All requests are now saying 503.
noticed the 503 errors for 2.5 pro model as well
Unfortunately there are several existing threads with the same issue, all of which are being ignored 
Hi
Were all request for Gemini 2.5 Flash/3.0 Flash met with a 503 error. ?
Was this during a specific timeframe during the day or was this some thing which you experienced throughout the day ?
I would also recommend implementing a exponential backoff to help mitigate the issue caused due to 503 errors
All day, every day for about the last week. Several have tried the exponential backoff with little success due to the latency caused by having to try 6 or 7 times before succeeding.
Also I’ve seen some reports of users being charged for the failed requests, though I haven’t confirmed this myself.
I’m in 29 Apr 2026 and same problem with my production application that i causing problems with my users in MicroChat1.0, Albanian’s first chatbot founded and developed by me