Amazon Bedrock adds advanced prompt optimization and explicit prompt caching features
Feature update Provisional 92% confidence first seen
Amazon Bedrock introduced two new capabilities for developers: Advanced Prompt Optimization, which automatically optimizes prompts across multiple models using reinforcement learning feedback, and explicit prompt caching for OpenAI GPT-5.6 models with 90% billing discounts for cached content. These features aim to streamline prompt engineering workflows and reduce inference costs for large-scale AI applications.