Tune caching for your application:
Tune caching for your application:
– Choose the prompt prefix to reuse via explicit cache breakpoints.
– Adjust reasoning effort and tool availability while preserving the cached context.
– Warm up shared context so responses start faster.