Request Standard (Regional) deployment for gpt-5.x in Australia East - healthcare workload

Joe 0 Reputation points
2026-07-20T05:04:08.68+00:00

We operate a healthcare application hosted in Australia East:

We need a pay-as-you-go model deployment with Standard (Regional) so that prompts and completions are processed in Australia East, per our compliance requirements.

Current situation

  1. Standard (Regional) in Australia East appears limited to gpt-4o and gpt-4.1-mini, both Deprecated — new deployments blocked (ServiceModelDeprecating).
  2. gpt-5-mini, gpt-5.1, and gpt-5.4-mini are only available as Global Standard, Data Zone Standard (Foundry may route to US), or Provisioned Throughput — none meet Australia-only inference at acceptable cost.
  3. We cannot deploy gpt-4o Standard or gpt-4.1-mini Standard despite the UI showing Standard for Australia East.
  4. US regions reportedly got gpt-5.1 on Standard as a gpt-4o replacement; Australia East has not.

What we need

  1. When will gpt-5.1 and/or gpt-5.4-mini be available for Standard (Regional) in Australia East?
  2. Is there an existing-customer / allowlist path to deploy gpt-4o Standard or gpt-4.1-mini Standard on our subscription until a gpt-5 successor ships in-region?
  3. Recommended compliant path for a healthcare app needing: vision/PDF extraction, json_schema output, Regional Standard (AU East inference), pay-per-token (not PTU).

Use case: File Upload - PDFs/images → Chat Completions with vision + strict JSON schema. Data is PHI under Australian privacy requirements.

We need a pay-as-you-go model deployment with Standard (Regional) so that prompts and completions are processed in Australia East, per our compliance requirements.

Not asking for: Provisioned Throughput (~$50/hr); Global Standard unless no regional option exists;

Foundry Models
Foundry Models

A catalog of AI models in Microsoft Foundry that you can discover, compare, and deploy using Azure’s built‑in tools for evaluation, fine‑tuning, and inference


Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.