AWS has announced that three OpenAI GPT-5.6 models—Sol, Terra, and Luna—are now generally available on Amazon Bedrock. The accompanying guidance covers how to choose among the models, run inference through the Responses API on the bedrock-mantle endpoint, and use prompt caching to reduce cost. It also describes connecting the OpenAI Codex coding agent and planning for quotas and scaling.
Why it matters
Making these models available through Bedrock gives users access to the GPT-5.6 family within AWS’s managed environment, along with tools for cost control via prompt caching and considerations for quotas and scaling.
Who should care
Developers and organizations building on Amazon Bedrock who want to use the GPT-5.6 models, integrate the Codex coding agent, or manage inference cost and scaling.