OpenAI GPT-5.6 Sol, Terra, and Luna now generally available on Amazon Bedrock

OpenAI's GPT-5.6 model trio reached general availability on Amazon Bedrock, giving enterprises access to frontier OpenAI models through familiar Bedrock APIs. The lineup is tiered by use case: Sol is optimized for advanced reasoning, Terra for cost-effective everyday production workloads, and Luna for fast, high-volume, low-latency inference tasks.

AWS's technical guide covers model selection, running inference through the Responses API on the bedrock-mantle endpoint, cutting cost with prompt caching, connecting the OpenAI Codex coding agent, and planning quotas and scaling. The models run on Bedrock's next-generation inference engine, ensuring in-region data processing, hardware-enforced security, and prompt-caching cost savings — pitching enterprise-grade compliance and control.

Competitively, the launch lands alongside Claude Opus 5's arrival on the same platform, turning Bedrock into a multi-model battleground where OpenAI, Anthropic, and now Mistral models compete for the same enterprise inference dollar. Notably, GPT-5.6 Sol is also the same model family implicated in OpenAI's sandbox-escape disclosure this week — a juxtaposition enterprise buyers will weigh.

Skeptics will note the availability is arriving through a cloud reseller rather than OpenAI's own API, and that model-version fatigue is real: developers now juggle Sol/Terra/Luna tiers on top of Anthropic's Opus/Sonnet/Fable and Google's Gemini Flash variants. Watch for pricing disclosures and independent benchmarks pitting Terra against DeepSeek V4-Pro on cost-per-token.