AWS customers in five Asia Pacific Regions can call GPT-6 Sol, Luna, and Astra through inference profiles that route to available capacity.

Amazon Bedrock now offers OpenAI’s GPT-6 Sol, GPT-6 Luna, and GPT-6 Astra models through global cross-Region inference for customers in Thailand, Malaysia, Singapore, Indonesia, and Taiwan. Developers pass a global cross-Region inference profile ID instead of a model ID on the bedrock-runtime endpoint, and Amazon Bedrock selects the destination Region that processes each request.

The GPT-6 family spans three models with different reasoning effort levels. GPT-6 Sol and GPT-6 Luna accept none, low, medium, high, xhigh, and max, while GPT-6 Astra accepts every level except none. All three take text and image input, client-side tool use, and prompt caching. AWS keeps billing, quota tracking, CloudTrail, and CloudWatch in the source Region.

Teams that hit throttling on single-Region inference can spread load across commercial AWS Regions without changing their monitoring or governance setup. Global profiles route requests to Regions where capacity exists, which helps operators run steady high-volume work and bursty reasoning jobs from the same source Region. Quota tracking stays local, so capacity planning keeps a single anchor point.

AWS documents inference profile IDs, IAM and service control policy configuration, and quota limits for teams moving production traffic. Monitoring stays tied to the source Region, and teams can invoke the profiles through the Converse API, OpenAI-compatible APIs, and coding clients. Broader destination Region coverage may follow as AWS expands capacity.

What matters

• Amazon Bedrock now serves OpenAI GPT-6 Sol, Luna, and Astra through global cross-Region inference profiles.
• Teams in five Asia Pacific markets can bypass local capacity limits when demand spikes.
• Operators should check inference profile IDs and quotas before shifting workloads onto the profiles.

Why it matters

Operators should check inference profile IDs and quotas before shifting workloads onto the profiles.

This GenAI News article was prepared in original wording using reporting and materials published by AWS Machine Learning Blog. Source reference: https://aws.amazon.com/blogs/machine-learning/accessing-openai-gpt-models-on-amazon-bedrock-with-global-cross-region-inference-in-thailand-malaysia-singapore-indonesia-and-taiwan/.

Drafted by the GenAI News review pipeline.