OpenAI GPT-5.6 Sol, Terra, Luna launch on AWS Bedrock
Three-tier model family offers 272K tokens and reasoning effort controls.
OpenAI has made its GPT-5.6 model family—Sol, Terra, Luna—generally available on Amazon Bedrock, giving AWS customers native access to frontier AI through their existing cloud infrastructure. The three models cover distinct use cases: Sol for autonomous coding agents, security research, and deep multi-step reasoning; Terra for general-purpose production workloads balancing reasoning, performance, and cost; and Luna for high-volume, latency-sensitive tasks like classification and summarization. All three accept text and image input, output text, and offer a 272,000-token context window with adjustable reasoning effort (none, low, medium, high, xhigh, max).
Developers can call these models using the OpenAI Responses API on the bedrock-mantle endpoint (e.g., https://bedrock-mantle.us-east-1.api.aws/openai/v1/responses), compatible with OpenAI Python and TypeScript SDKs. Pricing matches OpenAI first-party rates, and usage reduces AWS committed spend. Security features include IAM policies, VPC isolation, CloudTrail logging, and in-region inference for data residency. Prompts and completions are not used for training and are not shared with OpenAI unless you opt in. The models are available in US East (N. Virginia, Ohio) and US West (Oregon).
- Sol, Terra, and Luna cover autonomous reasoning, general production, and low-latency inference respectively, all with 272K-token context windows.
- Accessed via the bedrock-mantle endpoint using the OpenAI Responses API; pricing matches OpenAI direct rates and counts toward AWS commitments.
- Security includes IAM policies, VPC, CloudTrail, and in-region inference; data is not shared with OpenAI by default.
Why It Matters
AWS teams get frontier OpenAI models with no infrastructure ops, full cloud security, and flexible cost controls.