Plateforme et infrastructure de cloud computing pour la génération, le déploiement et la gestion d’applications et de services à travers un réseau mondial de centres de données gérés par Microsoft.
GPT-5.6 Sol unavailable on Microsoft for Startups subscription – Free Tier and 0 TPM in every region
I am using a verified Microsoft for Startups Azure subscription with active Azure credits.
We are preparing the production deployment of a B2B SaaS application and need to use:
- Model:
gpt-5.6-sol - Version:
2026-07-09 - Platform: Microsoft Foundry / Azure OpenAI
- Intended API: Responses API
The model is visible in the Microsoft Foundry model catalog, but it cannot be deployed because the subscription currently has 0 TPM / no deployable capacity for GPT-5.6 Sol.
The Foundry resource itself is healthy and located in France Central.
What I have already verified
1. Subscription quota tier
I checked the subscription quota tier directly through the Azure Management API from Azure Cloud Shell.
Command:
SUBSCRIPTION_ID=$(az account show --query id -o tsv)
az rest --method get \
--url "https://management.azure.com/subscriptions/$SUBSCRIPTION_ID/providers/Microsoft.CognitiveServices/quotaTiers?api-version=2025-10-01-preview" \
--query 'value[].properties.{Tier:currentTierName,Assigned:assignmentDate,UpgradePolicy:tierUpgradePolicy}' \
-o table
Result:
Tier Assigned UpgradePolicy
--------- ---------------------------- ----------------------
Free Tier 2026-09-01T20:00:53Z OnceUpgradeIsAvailable
So the subscription is currently assigned to:
currentTierName: Free Tier
tierUpgradePolicy: OnceUpgradeIsAvailable
2. GPT-5.6 Sol capacity across Azure regions
I then queried the Azure Cognitive Services modelCapacities API specifically for GPT-5.6 Sol.
Command:
az rest --method get \
--url "https://management.azure.com/subscriptions/$SUBSCRIPTION_ID/providers/Microsoft.CognitiveServices/modelCapacities?api-version=2024-10-01&modelFormat=OpenAI&modelName=gpt-5.6-sol&modelVersion=2026-07-09" \
--query 'value[].{Region:location,Type:properties.skuName,Available:properties.availableCapacity}' \
-o table
The API returned GPT-5.6 Sol entries across many Azure regions, but Available = 0 for every region and deployment type returned for this subscription.
Examples:
Region Type Available
------------------ -------------------------- ---------
australiaeast GlobalStandard 0
australiaeast DataZoneStandard 0
westeurope GlobalStandard 0
westeurope DataZoneStandard 0
eastus GlobalStandard 0
eastus DataZoneStandard 0
eastus2 GlobalStandard 0
eastus2 DataZoneStandard 0
uksouth GlobalStandard 0
francecentral GlobalStandard 0
francecentral DataZoneStandard 0
swedencentral GlobalStandard 0
swedencentral DataZoneStandard 0
germanywestcentral GlobalStandard 0
germanywestcentral DataZoneStandard 0
italynorth GlobalStandard 0
italynorth DataZoneStandard 0
spaincentral GlobalStandard 0
spaincentral DataZoneStandard 0
centralus GlobalStandard 0
centralus DataZoneStandard 0
canadacentral GlobalStandard 0
japaneast GlobalStandard 0
southindia GlobalStandard 0
The complete API response showed the same Available = 0 situation across all other regions exposed to this subscription.
This therefore does not appear to be only a France Central regional issue.
3. Official quota request already submitted
I also submitted an official Azure OpenAI quota request for:
Model: GPT-5.6 Sol
Deployment type: Data Zone Standard
Pool: EUR
The request was rejected with the following message:
The region you requested a quota increase in is not available to expand capacity at this time.
The response suggested trying additional regions.
However, the Azure modelCapacities API currently reports 0 available capacity for GPT-5.6 Sol in every region and pool exposed to this subscription, including Global Standard.
The bootstrap problem
This creates a difficult situation:
- The Microsoft for Startups subscription has 0 initial GPT-5.6 Sol TPM.
- Therefore GPT-5.6 Sol cannot be deployed.
- Because it cannot be deployed, we cannot generate GPT-5.6 Sol production traffic.
- Quota requests indicate that priority may be given to customers already consuming their existing quota.
- But our existing quota is zero.
Azure credits themselves are active and other Azure production/pre-production workloads are already generating real Azure consumption.
The problem therefore appears to be related to the initial Azure OpenAI quota tier / entitlement, rather than billing or Azure credit availability.
What I am trying to determine
Is an initial Tier 1 / GPT-5.6 Sol entitlement required before a Microsoft for Startups subscription can deploy this model?
If so, what is the correct process to obtain that initial entitlement when the subscription currently reports:
currentTierName = Free Tier
and:
gpt-5.6-sol availableCapacity = 0
across every region?
We are flexible regarding deployment location and deployment type.
We would be happy to use:
- Data Zone Standard EUR, preferably;
- or Global Standard in any region with available capacity.
We would also accept a small initial TPM allocation.
Our immediate goal is simply to deploy GPT-5.6 Sol, begin generating legitimate production traffic, and progressively scale the quota according to actual usage.
The workload is a real production B2B SaaS application using GPT-5.6 Sol for complex automotive diagnostic reasoning, long-context analysis, image understanding, Responses API workflows and web research.
Our production migration is planned within approximately two weeks.
Any clarification from the Azure OpenAI / Microsoft Foundry quota team, or experience from another Microsoft for Startups customer who encountered the same situation, would be greatly appreciated.