Introducing explicit prompt caching for OpenAI GPT-5.6 models on Amazon Bedrock
TL;DR
OpenAI GPT-5.6 Sol, Terra, and Luna are now generally available on Amazon Bedrock, along with explicit prompt caching that gives you precise control over which parts of your prompt are cached and reused. Learn how to get started, set up explicit caching, and migrate existing GPT workloads to reduce inference cost.
Nauti's Take
Explicit prompt caching is the practical part of this news: deciding yourself which parts of a prompt get cached can cut real cost on long system prompts instead of hoping a black-box heuristic gets it right. The limit shows up with dynamic prompts and sensitive data, where stale context turns into actual errors.
Teams on Bedrock get value from testing with realistic repetition rates, measured on hit rate and cost per request.
Sources
26.6.26