Hello,
The service seems overloaded for “gemini-2.5-flash-preview-04-17” model. I tried in the past hours and the model was responding without an issue for small size input (1k-2k tokens). But for bigger input (50k tokens) I am getting 500 Internal Error [except for only once when the model worked fine only during 10 minutes] . If the service is overloaded and there are issues thank you for updating the status page:
1 Like
After testing it seems the issue is with the way the model handles structured output and functions call. The solution could be in adjusting the prompt to avoid model confusion
Hey @rami_kan ,
I tried with more than 50k tokens with “gemini-2.5-flash-preview-04-17” model. It’s working fine on my end.
This appears to be a temporary issue.