Understand this story
On November 16, 2026, AWS Machine Learning announced that Amazon Bedrock's open-weight models now support explicit prompt caching. This feature can be used together with automatic implicit caching to reuse the already-computed state of the same prompt prefix, thereby reducing input costs and response latency. In terms of billing, cache reads are billed at a price lower than standard input, while for some models cache writes are priced higher; both modes are used through the Chat Completions API, and the specific supported modes, minimum prefix length, and TTL vary by model. There are currently no further developments.
AI-generated from public reports
Latest development · Report
Enable prompt caching for open-weight models on Amazon BedrockDevelopments
1 confirmed developments
Amazon Bedrock open-weight models support explicit prompt caching
Source:AWS Machine Learning
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 16
AWS Machine LearningFirst-party
Enable prompt caching for open-weight models on Amazon Bedrock
Open-weight models on Amazon Bedrock now support explicit prompt caching, which can be used together with automatically applied implicit caching to reuse the already-computed state of the same prompt prefix, thereby reducing input costs and response latency. Cache reads are billed at a lower price than standard input, and for some models cache writes are priced higher; both modes are used through the Chat Completions API, and the specific supported modes, minimum prefix length, and TTL vary by model.