Prompt caching with Claude. Caching lets you instantly fine-tune model responses with longer and more instructive prompts—all while reducing costs by up to 90%. Available in beta on the Anthropic API today.

→ View original post on X — @anthropicai