What would you like to see?

It would be helpful to support prompt caching for Claude 3.5 Sonnet. This not only helps cut down on LLM costs, but also latency.

https://docs.anthropic.com/en/docs/build-with-claude/prompt-caching

This can be done by adding an extra header, as an example...

with anthropic_client.beta.prompt_caching.messages.stream( system=system_prompt, model="claude-3-5-sonnet-20240620", max_tokens=8000, temperature=0, messages=messages, extra_headers={"anthropic-beta": "prompt-caching-2024-07-31"} ) as stream: