Introducing cross-Region inference for OpenAI GPT-5.6 models on Amazon Bedrock

Amazon Bedrock now offers OpenAI GPT-5.6 models with cross-Region inference (CRIS) capabilities, enabling requests to be routed across multiple AWS Regions for improved capacity and performance.

• Three GPT-5.6 variants (Sol, Terra, Luna) support geographic and global inference profiles for flexible routing
• Geographic profiles keep processing within specific geographies (e.g., US); global profiles route across all supported commercial AWS Regions
• Models accept text and image inputs, support 1 million token context window, reasoning mode, tool calling, and prompt caching
• Compatible with OpenAI Responses API, Chat Completions API, and Amazon Bedrock Converse API with streaming support
• Billing, quotas, and logging remain in source Region regardless of destination Region processing
• IAM policies control access to inference profiles with separate permissions for geographic and global routing
• Prompt caching reduces costs and latency by reusing cached prefixes across requests
• Output tokens consume quota at 10x burndown rate; cache read tokens don't count toward TPM quotas

GPT-5.6 on Amazon Bedrock enables scalable AI inference with flexible geographic constraints and unified billing across multiple AWS Regions.