google-gemini
Increased number of 503 errors - "The model is overloaded. Please try again later."
What happened?
When using gemini-cli in non-interactive mode, observing an increased number of the following error:
"error":{"message":"{\n \"error\": {\n \"code\": 503,\n \"message\": \"The model is overloaded. Please try again later.\",\n \"status\": \"UNAVAILABLE\",
We may be seeing similar reports in #7222, #7205, and #7222.
What did you expect to happen?
Clear handling of this error (do we retry, which model is the error coming from, etc.)
Client information
version: 0.2.1 OS: linux sandbox: none model: 2.5 pro
Login information
GEMINI API KEY
Anything else we need to know?
No response
Found possible duplicate issues:
- #4360
- #7205
If you believe this is not a duplicate, please remove the status/possible-duplicate label.
![gemini-cli[bot] avatar](https://avatars.githubusercontent.com/in/1473539?v=4 "Published by a pub.dev verified publisher"){kind=small}
I am also experiencing the same issue. It can't complete even one request.
Wonder if this is maybe caused by an uptick in model load of users due to 🍌
I have the same problem with 0.2.1 version, impossible work with 2.5-pro, I have to change to flash due to this problem
I have the same problem with 2.5-pro it's always overloaded !
Try updating to preview 4. Didn't have issues during a long sesh yesterday
Same issue here on v0.5.5, and the fact that's being around for a while doesn't look good. Prior to API key usage I never experienced it, while at the moment I see 61 of these errors using API key.
Same here, using API key connected to project with billing account, both latest stable and nightly Gemini CLI. Works flawlessly with the free tier personal auth, until quota is used up of course.
got new api error :
✕ [API Error: {"error":{"message":"{\n "error": {\n "code": 429,\n "message": "You exceeded your current quota, please check your plan and billing details. For more information on this error, head to: https://ai.google.dev/gemini-api/docs/rate-limits.",\n "status": "RESOURCE_EXHAUSTED"\n }\n}\n","code":429,"status":"Too Many Requests"}}] Please wait and try again later. To increase your limits, request a quota increase through AI Studio, or switch to another /auth method
login method : gemini api key
How does gemini-2.5-flash model compare to pro? Is it suitable for long running coding tasks? I'm constantly getting [API Error: The service is currently unavailable. (Status: UNAVAILABLE)] and it's getting frustrating
For tracking since this issue primarily focuses on the 503 errors and the fact that we haven't seen 503 errors a while, we will continue to keep an eye on this for the time being.
Installed just today and the very first query with gemini-2.5-pro returned:
✕ [API Error: got status: UNAVAILABLE. {"error":{"code":503,"message":"The model is
overloaded. Please try again later.","status":"UNAVAILABLE"}}]
$ date -u
Fri 3 Oct 2025 11:55:49 UTC
The next query went through just fine.
Is there any way to avoid this issue right now, while it's present? E.g., does switching auth from API key to OAuth login with Google Account fix it?
At this point this service is borderline a scam. It consistently drops every tuesday, wednesday, thursday and saturday. And on the days it doesn't it can still give this error from time to time, so even then it's not perfect.
It's a sinking ship unless they come up with something to finally fix it.
Lots of 503 400 etc lately. Esp the last week. And random quota is used. When it isn't.
Is anyone using Gemini in production given that this error keeps happening? How do you handle it? Fallback to OpenAI?