Accessing OpenAI GPT models on Amazon Bedrock with global cross-Region inference in Thailand, Malaysia, Singapore, Indonesia, and Taiwan
Amazon Web Services Traci Lim
AWS now lets Thailand, Malaysia, Singapore, Indonesia, and Taiwan reach OpenAI GPT-6 through Bedrock across Regions. That sidesteps local capacity limits without moving billing or logs out of the source Region.
Based on reporting by Amazon Web Services, Traci Lim — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
AWS is pushing OpenAI’s GPT-6 family deeper into Amazon Bedrock, and the main pitch is simple: if one Region runs hot, the request can go somewhere else. For teams in Thailand, Malaysia, Singapore, Indonesia, and Taiwan, that means access to GPT-6 Sol and GPT-6 Luna through global cross-Region inference from all five AWS Regions in this post, plus GPT-6 Astra from Asia Pacific (Singapore).
The setup is not just a model switch. You call Bedrock with a global cross-Region inference profile ID instead of a model ID, and Bedrock chooses the destination Region behind the scenes. The source Region is where the API call starts, and that is also where billing, quota tracking, CloudTrail, and CloudWatch stay. AWS says this is available through the Bedrock Runtime endpoint, and the post walks through the Converse API, the OpenAI-compatible APIs, and coding clients.
The three GPT-6 models split the work in different ways. Sol is aimed at stronger reasoning and coding on complex tasks, Luna is built for focused repeatable work at high volume, and Astra is positioned for complex reasoning, coding, and long-document work. All three take text and image input, client-side tool use, and prompt caching. The reasoning-effort controls are slightly different too: Sol and Luna accept none through max, while Astra skips none.
The bigger operational story is permissions. AWS spends a lot of time on IAM policies, service control policies, and Region activation because cross-Region inference does something awkward for enterprise setups: it sends a request without a Region name. That means an SCP Region allowlist can block it unless the exception is written carefully. AWS also points out that Singapore is enabled by default, while Jakarta, Malaysia, Thailand, and Taipei must be activated first.
There’s also a subtle warning baked into the post: profile availability and Region support can change as AWS adds more Regions and profiles. So the guidance is not just “use these profiles,” but “check the model card first, then wire up the policy, then watch the logs.”
My take — AI-written commentary, not fact-checked reporting
This is the kind of cloud feature that sounds boring until your favourite Region gets crowded. AWS is right to make the policy story front and centre, because cross-Region routing is where enterprise happiness goes to die in a spreadsheet. The real product here is not just GPT-6 access; it’s permissioning that doesn’t accidentally turn the whole org into a region-allowlist crime scene.
Read more about this at: Amazon Web Services