[FEATURE] Prompt caching support for all models #1140
Description
Problem Statement
This is a tracking ticket similar to Prompt caching support for LiteLLM but for all models.
As a user, I would like to leverage the SystemContent mechanism to apply prompt caching in all models.
Proposed Solution
No response
Use Case
To improve latency and reduce cost, Strands should support Prompt caching. As an abstraction layer, this should be done in a provider-agnostic way by using the SystemContentBlock.
Alternatives Solutions
No response
Additional Context
No response