AI Reports · Amazon · 2026-10-26 · Blog / commentary
Groundhog Day for AI agents: Prompt caching patterns for using OpenAI models on Amazon Bedrock
Every agent step makes the model re-read its entire context, including tool definitions, instructions, and history, which raises input cost and latency. Prompt caching for OpenAI models on Amazon Bedrock stores the input prompt and reuses it on later requests, cutting cost and latency without resending the same context each turn. It also helps agents stay responsive and consistent as they gain new tools and spawn helpers.