Zero quota for GPT-6 Luna and Claude Sonnet 5.5 while GPT-6 Sol works: request human eligibility review

David Levy 0 Reputation points
2026-09-28T20:13:32.58+00:00

I am requesting help from Microsoft Foundry quota/support staff with a subscription-level eligibility review and starter quota allocation. Azure Billing and Subscription Management Support directed my existing quota case to Microsoft Q&A because the subscription has a Developer support plan. Please route this to Priority Community Support / the appropriate Foundry quota team.

Current control-plane checks on September 28, 2026:

  • GPT-6 Sol (2026-09-22), Global Standard: quota limit 100 units, 100 allocated, and a successful serving deployment (100,000 TPM). The application is operating on this subscription.
  • GPT-6 Luna (2026-09-22): Global Standard and Data Zone Standard limits are zero, and all 44 returned model-capacity records across supported regions report zero. Luna previously had a working deployment on the source subscription before migration; that source subscription has since been canceled.
  • Claude Sonnet 5.5 appears in the existing Foundry resource model catalog as Anthropic versions 1 and 2, but quota is zero for both hosting versions. All 9 version-1 and 17 version-2 capacity records report zero.
  • The official quota-increase form at https://aka.ms/oai/stuquotarequest does not currently list Sonnet 5.5 under either Anthropic hosting menu. It lists Sonnet 5, but I did not submit a request for a different model.

Requested allocation: 50,000 TPM each for GPT-6 Luna and Claude Sonnet 5.5, or the smallest usable starter allocation you can provide. One usable Global Standard region per model is sufficient; East US is preferred, but another supported US region is acceptable, and either Sonnet hosting version is acceptable. These are alternative placements, not additive regional allocations.

Use case: an existing construction-document application, with bounded development/evaluation of source-grounded text and visual reasoning, low concurrency, and explicit request/time/output bounds. Sol is already serving; Luna would provide a lower-cost route, and Sonnet would enable a separately evaluated alternative. We cannot consume Luna or Sonnet quota while the limit is zero.

Could Microsoft please (1) identify the entitlement/allocation path that enabled Sol and review whether it can enable Luna, (2) provide a valid Sonnet 5.5 quota-request route while the form menu is missing the model, and (3) escalate for a human review rather than repeat an automated regional-capacity rejection?

I can provide the subscription ID, existing support case number, and additional evidence through an official private Microsoft support channel. Please preserve the existing Sol deployment and quota. This request is for pay-as-you-go standard inference only; no provisioned throughput, paid support upgrade, or new subscription is requested.

Additional eligibility evidence: Azure emailed that this subscription is eligible to move from Free Tier to Tier 1 automatically. The quotaTiers API still shows Free Tier, next tier Tier 1, upgradeAvailabilityStatus Available, tierUpgradePolicy OnceUpgradeIsAvailable, and upgradeApplicableDate 2026-09-29T21:01:19.8788872Z. Please confirm whether that pending tier upgrade will enable these models, or whether separate model allocations are required. I am requesting quota review now without opting out of the scheduled upgrade.

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

0 comments No comments

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.