I’m using gemini-2.0-flash in my LangGraph agent on one of the nodes. It was working fine last Friday, but this morning it fails to call the tools (it never triggers them). I’m using temperature=0.
This node has predictive behavior, and in the past, even if tool calling fails once or twice, it eventually succeeds. However, today it doesn’t call the tools at all.
Has there been any update or quantization to the model over the weekend? Is anyone else experiencing similar issues?
Hi @codeonym
Welcome to the AI Forum!!!
Thank you for bringing this to our attention.
Could you please try to use our latest model and let us know if the issue still persist?
Thank you!
Hi @Shivam_Singh2 , thanks for the reply ! Just to clarify a few points. I’m currently not experiencing any issues with other models regarding tool calling. The reason I initially preferred using 2.0 is because it’s integrated within a routing node - it’s simpler in terms of complexity and delivers much faster responses (latency).
However, I’ve now switched to 2.5-flash-lite-preview-09-2025, assuming it provides better performance than 2.0, and so far it’s working well.