Flash seems to be fine, whether model is simply inserted for chat completion or used for another pipeline in my app such as translation. But in chat, pro times out, and in other use cases it consumes about 10k tokens then gives an error (either 500 internal server error or ‘markdown.split is not a function’ - see image). Consistent since enabling Gemini via Langchain wrapper for the last 3 days.
Context length is already controlled variable.
Hello,
Welcome to the Forum!
This appears to be an internal error and it should be resolved after some time. Could you please try again and let me know if it works?
Hey all,
I also have the same Problem. Flash is working via API, but Pro is producing a 500 Internal Error..
If this issue still persists, we need to analyze this issue. Could you please share when this first occurred and provide the full request payload you sent when the error occurred?
Seems to be working now, can close this