I use gemini-1.5-pro-preview-0409 model via API, get error: Socket closed

[adorehb]

Using the gemini-1.5-pro-preview-0409 model through the API, but when the number of tokens is high （10k）, the server directly reports 'Socket closed'.

[nceniza]

You must be reaching either the input or the output token limit for 1.5 version:

Input token limit	1048576
Output token limit	8192

The model is in preview state, it may lack support and is unstable vs GA models available at the moment.

Ref:

https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versioning#preview-version