AI article
Prompt Caching: Why cache_control Writes But Never Reads
Community description: Prompt caching can raise your bill: why cache_control writes on every request and never reads, and how to fix breakpoint placement.
Dev.to | Sep 12, 2026 | jidonglab
Automated excerpt
I turned on prompt caching for an agent loop that resends a 12K-token system prompt on every turn. Render order is tools → system → messages. Why does prompt caching never read from the cache? Because your cache_control breakpoint is placed after content that changes every request, so each request writes a brand-new entry that nothing will ever match again. What makes prompt caching miss on byte-identical requests?
Selected automatically from source text; not independently written or fact-checked. Read the original for full context.