Empty responses seem to have lessened a bit, but now a new issue has appeared. I keep getting this error when I try to use gemini-2.5-pro.
same issue with alot of users
Facing the same issue.
The empty responses have been fixed by the internal team and the changes have been rolled out. This “Model Overload” might be a temporary due to the rollout, Please wait for some time and try again.
let us know if this issue persists over a long period of time
i am getting so many API errors including 503 service unreachable, 404 not found and 503 server overload, my project is about data analysis and it kept failing from two days for almost 18 hours a day, many times my data analyst agent couldn’t even call the LLM even after 3-5 retries…I have started to design my project when the free tier have 5RPM(req per min) but just few days ago google has drastically reduced to on 2RPM which has caused huge disruptions during my proj evaluation so I have paid for tier-1 and still getting these errors. My semester grade
depends upon this project
Facing the same issue.
Yep, same issue ehre
facing same issue on paid tier from 3 days now
The same issue, return code 503
The same issue, return code 503
Why is it that these type of issues don’t happen in AI Studio but are frequently happening in the API? I’m sure we are going to get these when Gemini 3 comes out too…
Try switching to other auth method ----> /auth —> google account
I set up an uptime monitor to ping every 5 minutes, and in the last 3 hours, there have been no successful requests.
2 Likes
Really fixed? Do you mean that continuous occurrence of the internal server error is normal?
This has been going on for months.
How are we to build stable services for our users if the models are constantly overloaded?
If Google can’t get enough servers to deploy these models, who can?