Better Prompt Caching for GPT-6
OpenAI, Tuesday, September 22nd, 2026
OpenAI improved GPT-6 prompt caching with higher default hit rates and new dashboard and diagnostic tools for developers.
OpenAI upgraded prompt caching for the GPT-6 model family, delivering higher cache hit rates by default and extending cache discounts to eligible shared prefixes reused within a 30-minute window, with discounts of up to 90% on cached input tokens.
A new Prompt Caching Dashboard shows how much input is served from cache and how it changes over time, while a diagnostics tool compares requests to identify what caused an unexpected cache miss.
Explicit cache breakpoints let developers choose which prompt prefixes to cache, supported by a refreshed prompt caching guide.