Function Calling Multimodal Live API

Google AI Studio
1

Hi everyone,

I’m using the websockets and audio streaming for the API and there are times when the model makes a function call but after returning the response to the model it seems like the model didn’t even process it. Is anyone else facing this issue?

Thanks

4 Likes
3

I’ve been encountering the same issue. I’m trying to integrate a knowledge base and have a Python API that analyzes queries and retrieves similar content using Sentence Transformers. While the API returns responses correctly, the model doesn’t seem to reference them at all.

5

Hi @blue1 Apologies for late response
Hope this issue is resolved by now . If not could you specify Gemini model are you using?Could you also please try to use latest models and let us know if the issue still persist? Thank you!

1 Like
7

Hi @Hemanth_T_S @blue1,

The gemini-2.5-flash-native-audio-preview-09-2025 model offers significantly improved function calling. Please reattempt your task with this new model and let us know your thoughts.

Thank you!