OpenAI says GPT-6 prompt caching now provides higher default cache hit rates and discounts for eligible shared prefixes reused within a 30-minute window. Developers can use a prompt caching dashboard and diagnostics to inspect hit rates, investigate misses, and optimize reusable context with explicit breakpoints and stable tool definitions. OpenAI says cached input-token discounts can reach 90%, while actual savings depend on request structure and reuse.

Editor's note

This summary is based on the primary source. Open the original for full context, data, and qualifying details.