Getting 500s With Experimental Models

Gemini API
1

I’m getting this error consistently for the past few days:
line 102, in raise_for_response raise ServerError(status_code, response)google.genai.errors.ServerError: 500 INTERNAL. {'error': {'code': 500, 'message': 'An internal error has occurred. Please retry or report in https://developers.generativeai.google/guide/troubleshooting', 'status': 'INTERNAL'}}(env)

I traced the error down to the model I’m using. Using gemini-1.5-pro, it works. Using gemini-2.0-pro-exp-02-05, it gives the 500 error. This program uses large files, but the files are under the token limit. If I use AI Studio with the same prompts and files, it works.

I can provide the files and code privately to a Googler to replicate it.

1 Like
3

Hi @Jesse_Anderson, If possible could you please share the files and the prompt you are using to replicate the issue. Thank You.

4

I tried again, and I’m not getting the 500 error anymore.

A different question: how do I get more quota for experimental models?

5

Hey Jesse, we don’t offer higher rate limits for experimental models right now, but we are actively working to enable this! See current limits here: Rate limits  |  Gemini API  |  Google AI for Developers

2 Likes