Persistent An internal error 500

Gemini API
,
1

In the last 10 days Gemini EXP models became very unreliable, and report the following error via API:

InternalServerError: 500 POST https://generativelanguage.googleapis.com/v1beta/models/gemini-exp-1114:generateContent?%24alt=json%3Benum-encoding%3Dint: An internal error has occurred. Please retry or report in https://developers.generativeai.google/guide/troubleshooting

the following models are affected:

  • Gemini EXP 1114
  • Gemini EXP 1121
  • Gemini EXP 1206

the following models are NOT affected:

  • Gemini EXP Flash 2-0

It all started December 24/25 and now happens every 2-3 days randomly and lasts for 4-20h. And at the moment of writing, again, Gemini EXP models do not work

  • Tried to use legacy SDK, new SDK, openAI SDK
  • Tried to create a new API key
  • Tried to remove system prompt / samplers / settings
  • Tried to use various prompts
    …and nothing worked

There are already multiple threads in forum but I decided to create a new one for clarity and attention.

2 Likes
2

Just here to co-sign. Same experience, same symptoms. I would even say it started way back during the Friday of the Flash 2.0 launch. 500 Internal errors for any queries over ~100k tokens.

2 Likes
5

not only this issue was NOT resolved in all those days and still occurs frequently - but I also learned just now, that those errors 500 COUNT towards QUOTAS

image
image561×628 17.1 KB

so yes right now I have 57/100 RPD and I haven’t got a single answer - I only get error 500. I actually was checking whether EXP models were back every 5 minutes. and it was enough for kill half of my quota

is this a joke?