rfc(anthropic): support cache_control as a kwarg

[ccurme]

from langchain_anthropic import ChatAnthropic

llm = ChatAnthropic(model="claude-3-5-haiku-latest")
caching_llm = llm.bind(cache_control={"type": "ephemeral"})

caching_llm.invoke(
    [
        HumanMessage("..."),
        AIMessage("..."),
        HumanMessage("..."),  # <-- final message / content block gets cache annotation
    ]
)

Potentially useful given's Anthropic's incremental caching capabilities:

During each turn, we mark the final block of the final message with cache_control so the conversation can be incrementally cached. The system will automatically lookup and use the longest previously cached prefix for follow-up messages.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

1 Skipped Deployment

Name	Status	Preview	Comments	Updated (UTC)
langchain	⬜️ Ignored (Inspect)			Jun 6, 2025 10:12pm

[codspeed-hq]

CodSpeed Walltime Performance Report

Merging #31523 will not alter performance

_{Comparing cc/anthropic_cache_kwarg (9d8342b) with master (5187817)}

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

Summary

✅ 13 untouched benchmarks

[codspeed-hq]

CodSpeed Instrumentation Performance Report

Merging #31523 will not alter performance

_{Comparing cc/anthropic_cache_kwarg (9d8342b) with master (5187817)}

Summary

✅ 13 untouched benchmarks

[sydney-runkle]

Seems reasonable, anything blocking merge here?

Have we done integration testing with this diff?