Claude Prompt Caching in 2026: The 5-Minute TTL Change That's Costing You Money

If you're running Claude API workloads and haven't checked your caching bill lately, you're in for a surprise. Anthropic quietly changed the prompt cache TTL from 60 minutes down to 5 minutes in early 2026. For many production workloads, this single change increased effective API costs by 30–60%. Here's what changed, who it hits hardest, and how to architect around it. What Is Prompt Caching? Claude's prompt caching lets you cache expensive prefill tokens (system prompts, long documents, tool de