AWS Bedrock adds cross-region inference for GPT-6 models across Southeast Asia
AWS ML Blog · rss · 2026-10-13
AWS announced that five Asia-Pacific regions (Singapore, Jakarta, Malaysia, Thailand, Taipei) can now call OpenAI's GPT-6 family via global cross-Region inference, avoiding throttling when single-region capacity runs out. GPT-6 Sol and GPT-6 Luna cover all five source regions; GPT-6 Astra is invoked from Singapore.
- Model positioning: Sol for heavy reasoning and coding, Luna for high-volume repeatable tasks, Astra for complex reasoning, coding, and long documents. All three accept text+image input, tool use, and prompt caching, with reasoning effort from none to max (Astra rejects none).
- How it works: pass a cross-region inference profile ID (e.g. global.openai.gpt-6-sol) in place of the model ID; billing, quotas, and CloudWatch stay in your source region. Supported via Converse API, OpenAI-compatible APIs (/openai/v1, SigV4 or API key), and coding clients, with full Python examples using boto3 and the OpenAI SDK with aws-bedrock-token-generator.
- Prerequisites: only the source region needs activation (Jakarta, Malaysia, Thailand, Taipei are opt-in), plus appropriate IAM permissions.
More from Infra
- TSMC September Revenue Jumps 54.6% YoY, Q3 Hits $46.7B on AI Demand — Beth_Kindig · 2026-10-13
- Open-source warmctl restores a warmed-up vLLM engine in ~37s on 8×H100 — kbesplv · 2026-10-13
- NVIDIA open-sources SMLIB, its NURBS-based CAD kernel, after 20-year partnership — Promptmethus · 2026-10-13
- STAPLE adds static sparse memory to pretrained LLMs with 50% less serving energy than MoE — BeidiChen · 2026-10-13
- OCP chatter: tighter scale-up/out compute domains could ease the race to ~500 GPUs per rack — BenBajarin · 2026-10-13
- NVIDIA releases NVFP4 quantized Qwen-Image with 12B-size weights on Hugging Face — TheZachMueller · 2026-10-13