Handling 429 / 503 errors from the Gemini API

Gemini API
1

If you’re experiencing 429 or 503 errors when accessing the Gemini API, please read on.

429 errors

429 (“Too Many Requests”) errors are usually triggered if you exceed one of our listed quotas:

  • Requests per minute (RPM)

  • Tokens per minute (TPM)

  • Requests per day (RPD)

First: check your usage & quota

  • In: AI Studio > Dashboard > API Keys > View usage (the bar chart symbol next to your project), check the Quota & Rate Limit tabs. Usage data is delayed ~15 mins.

  • Double check you’re using the API key for the correct project.

Still getting 429’s?

  • If you’re running into limitations from RPM or TPM limits, consider implementing backoff and retry with a long maximum delay (our SDKs support this through retry config).

  • If no rate limits are being breached, please add your details to this form & we’ll check it out.

503 errors

503 (“Service Unavailable”) errors are unrelated to your quota and indicate our services are temporarily overloaded and can’t service your request. These errors are more common during certain times of the day. We apologise for the impact of these errors & are working hard to provision more resources and reduce their occurrence.

To handle these errors, we strongly recommend implementing backoff and retry (our SDKs support this through retry config). Although your requirements will vary based on your desired user experience, you can try setting a low number of retries with a short timeout at first, and increase these if the issue isn’t resolved.

If applicable, you could also consider moving non-realtime jobs to our Batch API.

2

AI Studio is returning consistent 503 and 504 errors on 3.1 Pro Preview since the Build has been ‘fixed’ (though it was already slow prior, now it’s just down for me). 3.0 seems to go through.

3

Same issue for me.

3.1 pro is not responding, although 3.0 pro is.

6

Experiencing the same issue. What’s even worse is that important production projects for a client are currently down because of this…

7

I’m getting those errors because Google won’t upgrade me to the next spending tier

8

I’m seeing so many errors from the Gemini API that it’s not even funny. Using the gemini-2.5-flash-lite-latest, I get a lot of 503 errors, a lot of 400 errors (Cannot fetch content from the provided URL, but URL can be read by other services) and even more timeouts after 30 seconds. The average response-time is 10 seconds compared to GoogleCloudVision which took about 4 seconds. Overall the response-times of the service have multiplied by factor 10 due to retries and timeouts.

It’s saying a lot, that there is no response-time graph in AI studio.

9

I’ve been having this issue since last night (Attempt 1 failed: 429 RESOURCE_EXHAUSTED. {‘error’: {‘code’: 429, ‘message’: ‘Resource has been exhausted (e.g. check quota).’, ‘status’: ‘RESOURCE_EXHAUSTED’}}). I thought it was because of hitting the quota, so I stopped around 1am EST. I read that the quotas are reset at midnight PST, which would be 3am EST. I’m still having the issue today. Have folks found out what it might be and how long to wait if it really is about internal quota limits?

10

@Jon_Matthews Hi, I’m experiencing a critical issue that goes beyond normal rate limiting.

Paid Tier 1 with active billing, but all regional quotas are set to 0 — not just one region, every single region.

Error details:

quota_metric: generativelanguage.googleapis.com/generate_content_requests
quota_limit: GenerateContentRequestsPerMinutePerProjectPerRegion
quota_limit_value: "0"

Note: the metric does NOT contain the free_tier prefix, confirming paid tier recognition. However, the per-region quota is hard-locked at 0 across all regions.

In the Cloud Console Quotas page, every region shows “between 0 and 0” — I cannot self-service increase it. This is not a normal rate limit issue; it appears to be an erroneous producer override.

I’ve verified this by creating a new account with billing — works perfectly with normal Tier 1 quotas. The issue is isolated to this specific project.

Could someone from the team please look into removing the incorrect override on this project? Thank you.

11

It’s disgusting to read. For three months now, I haven’t been able to get a normal response from the model without interruptions or errors. And they’re telling me it’s my fault. I switched to a different model and am getting responses without errors. Get out of here as fast as you can.

12

100% getting 429 errors when my rate limit page (https://aistudio.google.com/rate-limit) is showing me nowhere near hitting any limits.

I’m using Geminin within the Agent window of Android Studio. It’s showing it linked to my paid account, with the ‘Gemini for business’ badge, and everything appearing to be hooked up. And yet every hour or so while trying to work it’ll just stop and say it’s exceeded quotas.

I signed up for the paid plan last week and have had these 429 errors constantly. Does not make me want to continue paying for it at all.

13

I can confirm that gemini is returning 429 while the limit is about 30k requests away as of two days ago.

14

I am getting 503 errors from Gemini 3.1 flash lite. No errors past few days, just started today.

15

get a 429 error when I do :

··• gemini -p “test”
Keychain initialization encountered an error: The name is not activatable
Using FileKeychain fallback for secure storage.
Loaded cached credentials.
Attempt 1 failed with status 429. Retrying with backoff… GaxiosError: Resource has bee
n exhausted (e.g. check quota).

been like this for 4 days, I am on Google Pro Ai package and this is not on. Can’t get help either from the Ai could platform. Is this a permanent quota restriction??

When I try to access my rate limites i get this message at the top of the screen: Project quota tier unavailable. Please contact your project administrator for assistance.

Google have messed up big on this.

17

Yes the gaslighting from Google on this is infuriating. As of this morning I’m still getting constant 500 and 503 errors with my paid API key on 4K requests. Over the weekend and yesterday, 2K was working fine - and as of this morning, 2K AND 1K requests are simply failing every time.

18

I cancelled my AI Ultra plan over this, as have many others recently. See this thread in which hundreds of paying customers are completely ignored on the Gemini CLI Github page for more details.

19

This issue is a plague on any user trying to work with Gemini in Antigravity. The application Google has made for us to use, that application is causing the issue. It doesn’t matter if you are on the free plan or paying $250/month, this issue comes up regardless.

It seems there are some very fundamental problems Google needs to be figuring out as this will inevitably leave to cancellation of all service

20

Yep, can’t even work thanks to this persistent issue.

[GeminiAPI] Stream request failed: 429 {
“error”: {
“code”: 429,
“message”: “No capacity available for model gemini-2.5-pro on the server”,
“status”: “RESOURCE_EXHAUSTED”,
“details”: [
{
“@type”: “type.googleapis.com/google.rpc.ErrorInfo”,
“reason”: “MODEL_CAPACITY_EXHAUSTED”,
“domain”: “cloudcode-pa.googleapis.com”,
“metadata”: {
“model”: “gemini-2.5-pro”
}
}
]
}
}

21

Still getting after latest Android Studio update. Using paid plan subscription. Intentionally NOT using the API Token, because that starting costing money outside of the subscription, which is ridiculous.

I’m now about to try Claude or OpenAI instead of Gemini as the model within Android Studio, which is a ridiculous state of affairs, you should 100% be able to easily use Google’s model within Google’s IDE to develop apps for Google’s mobile OS.

ALL I’m doing is using the agent with Android Studio, that’s my only use of Gemini, and yet I’m hitting this time and again.

POST https://cloudcode-pa.googleapis.com/v1internal:streamGenerateChat
{
“code”: 0,
“error”: {
“code”: 429,
“message”: “You have exhausted your capacity on this model. Your quota will reset after 49m30s.”,
“errors”: [
{
“message”: “You have exhausted your capacity on this model. Your quota will reset after 49m30s.”,
“domain”: “global”,
“reason”: “rateLimitExceeded”
}
],
“status”: “RESOURCE_EXHAUSTED”,
“details”: [
{
“@type”: “type.googleapis.com/google.rpc.ErrorInfo”,
“reason”: “QUOTA_EXHAUSTED”,
“domain”: “cloudcode-pa.googleapis.com”,
“metadata”: {
“uiMessage”: “true”,
“model”: “gemini-2.5-pro”,
“quotaResetDelay”: “49m30.938289688s”,
“quotaResetTimeStamp”: “2026-04-10T01:47:33Z”
}
},
{
“@type”: “type.googleapis.com/google.rpc.RetryInfo”,
“retryDelay”: “2970.938289688s”
}
]
}
})

22

Hi jade many error and a json leak and chain of thought leak and may other glitches