Gemini API returns 429 issue when using OpenAI compatible API

Gemini API
,
1

I have been using gemini google API with openAI format through this API

https://generativelanguage.googleapis.com/v1beta/openai/

But suddenly I’m getting this error

RateLimitError: Error code: 429 - [{'error': {'code': 429, 'message': 'Resource has been exhausted 
(e.g. check quota).', 'status': 'RESOURCE_EXHAUSTED'}}]

Even though I’m sending only one query at a time with small question "What is the capital of Egypt)

If I used gemini through the google sdk it works just fine. Did I hit the OpenAI format limit?

3

Yes, you did.
Wait a little bit or try again tomorrow.

5

Do you have any link or documentation on this, like what is the limit ?

6

We are experiencing the same issue.
We are on paid plan and we are not reaching anywhere close to the quota based on usage statistics.
When we request to /generate_content API, it does not return quota error. But when we request with stream=True, it returns different error.

8

I am also experiencing the same issue specifically when streaming the response. I am getting frequent 429 errors.

9

It looks like the action items are:

  • Google to provide documentation on the OpenAI format limit.
  • Google to investigate the 429 error issue with streaming responses.
1 Like
13

Hi google team, is there any update on this issue? We have also tried using the latest openai library version but the issue persists.
Using google-generativeai generate_content() with the same key works but the issue only occurs with OpenAI format (and it was working fine last week).