Quota shows as fully utilised for Australia East region for Standard deployment

Hello @Deepak Sharma

Thank you for reaching out to Microsoft Q&A.

Based on the behavior you described, it does appear that you may be encountering quota/capacity retention from previously deleted Standard deployments in Australia East.

Even though the deployments were deleted, Azure OpenAI Standard deployment capacity can sometimes remain reserved temporarily due to backend capacity allocation tracking and regional quota refresh cycles.

Here are a few checks and recommendations that should help clarify the situation:

1. Confirm why the deployment request is being denied

In the Azure AI Foundry portal: • Go to Azure AI Foundry → Management Center → Quota • Review the quota status for the affected model/region

If the portal reports: • “regional capacity unavailable” or • deployment denied due to capacity constraints,

this typically indicates that the Standard capacity pool for that model in Australia East is currently exhausted or still accounting for prior allocations.

1. Verify actual quota consumption

Please check the current usage values directly using Azure CLI or REST API.

Azure CLI:

az cognitiveservices usage list --location australiaeast

REST API:

az rest --method get \
--url "https://management.azure.com/subscriptions/{subscriptionId}/providers/Microsoft.CognitiveServices/modelCapacities?api-version=2024-10-01&modelFormat=OpenAI&modelName={yourModel}&modelVersion={yourVersion}&skuName=Standard"

Please review: currentValue, availableCapacity

This can help confirm whether capacity is still being counted against prior deployments.

1. Possible orphaned or delayed quota release

Standard deployment capacity pools can sometimes refresh on an internal allocation cycle (commonly up to ~30 days depending on backend reconciliation timing).

If the deleted deployments were removed recently, their reserved capacity may not yet have been fully released back into the regional pool.

If more than 30 days have passed, and quota/capacity still appears consumed, then this may indicate orphaned backend allocation state, in which case a backend investigation/support escalation may be required to clear the retained allocation.

1. Short-term workarounds

While the backend allocation refresh is being validated, you may consider:

Deploying in another nearby supported region (for example Australia Southeast) • Using Global Standard deployment type if your workload supports it, since Global Standard uses a separate capacity pool • Trying a smaller deployment size/capacity allocation temporarily

Please refer this

Resolve Quota Allocation Issues for Azure Regions: https://learn.microsoft.com/azure/ai-services/openai/how-to/quota

I Hope this helps. Do let me know if you have any further queries.

Thank you!