Model Settings

Live
Per-model defaults are applied when neither the request nor the matched routing rule specifies max_tokens or temperature.