GPT 5 azure AI model pricing not available for regional standard deployment

Shivangi Prasad 0 Reputation points
2026-11-06T10:35:41.4+00:00

We wish to deploy our resource in Regional standard deployment in Australia east for GPT 5 model. But a) there is no option in deployment dropdown to select "Standard" and

b) the official azure pricing documentation does not contain Regional standard price rates for GPT 5

Azure OpenAI in Foundry Models
0 comments No comments

1 answer

Sort by: Oldest
  1. AI answer

    2026-11-06T10:44:22.4233333+00:00

    GPT-5 version 2025-08-07 cannot currently be deployed as Regional Standard in Australia East; use Global Standard, or Regional Provisioned Managed when geography-based processing is mandatory.

    The missing Standard option and missing Regional Standard price are therefore consistent with the documented availability rather than a portal or quota problem. The current matrix lists GPT-5 2025-08-07 in Australia East for Global Standard and Regional Provisioned Managed, but does not confirm Regional Standard there.^1^

    Currently documented choices

    Requirement Deployment choice Billing/processing consequence
    Pay per token; global processing acceptable Global Standard (GlobalStandard) Available from Australia East, but inference can be processed in any supported Azure region.
    Geography-based processing is mandatory Regional Provisioned Managed (ProvisionedManaged) Documented for GPT-5 in Australia East; reserved PTU capacity rather than pay-per-token. Processing remains within the applicable Azure geography and can occur between regions in that geography.
    Asynchronous workload acceptable Global Batch (GlobalBatch) Available for GPT-5 in Australia East, but processing is global and has a target turnaround of 24 hours.
    APAC-wide processing acceptable and model change allowed Data Zone Standard with a listed successor, such as GPT-5.2 The current APAC table lists GPT-5.2, not GPT-5 2025-08-07. Validate the exact model/version in the live table before deployment.
    GPT-5 Regional Standard specifically required Not currently documented in Australia East No self-service configuration or quota increase can add an unsupported model/deployment combination.

    Microsoft documents that new deployment types generally arrive in this order: Global, then Data Zone, then geography-based. Geography-based availability arrives last, depends on capacity, and has no guaranteed availability date.^2^

    Do not interpret an Australia East Global Standard deployment as Australia-only processing: Australia East is where the resource/deployment is created, while Global Standard may route inference to other Azure regions.

    Pricing interpretation

    The pricing page lists GPT-5 2025-08-07 Global and Data Zone rows but no GPT-5 Regional row.^3^ Without an available Regional Standard deployment and corresponding regional meter, there is no applicable public GPT-5 Regional Standard price for this combination.

    For an independent meter check, query the public Azure Retail Prices API. This example follows pagination and restricts results to Australia East, GPT-5-family consumption meters, excluding global, Data Zone, batch, and provisioned SKUs:^4^

    import requests
    
    url = "https://prices.azure.com/api/retail/prices"
    params = {
        "$filter": "productName eq 'Azure OpenAI GPT5' "
                   "and priceType eq 'Consumption'",
        "api-version": "2023-01-01-preview",
    }
    items = []
    while url:
        data = requests.get(url, params=params, timeout=30).json()
        items += data.get("Items", [])
        url, params = data.get("NextPageLink"), {}
    
    for item in items:
        sku = item["skuName"].lower()
        excluded = ("global", "data zone", "batch", "prov")
        if item["armRegionName"] == "australiaeast" and not any(x in sku for x in excluded):
            print(item["skuName"], item["retailPrice"], item["unitOfMeasure"])
    

    Inspect any returned SKU name carefully for the exact model and meter direction. A retail meter’s existence alone does not establish that the model is deployable; the model/version/deployment-type/region availability matrix remains authoritative.

    If pay-per-token Regional Standard is a firm requirement, retain the existing architecture plan but open an Azure support request for roadmap or business-impact guidance. There is no published availability date to rely on.


    References

    1. Region availability for Foundry Models sold by Azure (standard)
    2. Deployment types for Microsoft Foundry Models
    3. Azure OpenAI Service - Pricing | Microsoft Azure
    4. Azure OpenAI frequently asked questions
    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

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.