Question

You are building a lightweight chat client with the Foundry SDK. Users complain that responses feel slow because the entire answer only appears after the model finishes generating. You want the client to display text progressively as it is produced, without changing the deployed model. What should you enable in your SDK call?

Accepted Answer

Streaming responses, so tokens are returned incrementally as they are generated

Answer

A higher max_tokens value to speed up total generation

Answer

A lower temperature setting to make output more deterministic

Answer

Function calling so the model offloads work to an external tool