Best LLM Inference Providers

Vendor profileChecked Nov 8, 2026

Microsoft Foundry (formerly Azure AI Foundry)

Microsoft Foundry on Microsoft Azure; the Azure AI Foundry product URL now serves the Microsoft Foundry page · azure.microsoft.com · Last checked:

  • Cloud provider platform
  • Serverless per-token API
  • Provisioned throughput (PTU)

Company status

Active. Microsoft Foundry is an Azure service and business line of Microsoft Corporation.

Sources 1, 2, 3, 4, 5 · checked 2026-11-08

Capabilities

  • Proprietary Azure OpenAI models and other Foundry Models sold by Azure (for example DeepSeek, Grok, and Microsoft AI models) plus partner models
  • v1 API works with the OpenAI client, including chat completions for models from other providers
  • Standard, Global, and DataZone deployment types that set where prompts are processed (for example, US or EU data zones)
  • Prompts and completions not used to train foundation models and not available to OpenAI or other model providers
  • Batch processing and fine-tuning for supported models

Sources 1, 2, 3, 4, 5 · checked 2026-11-08

Who it tends to fit

  • Teams on Azure that need Azure OpenAI models under Azure contracts, networking, and Entra ID
  • Buyers who need EU or US data-zone processing for models sold by Azure

Sources 1, 2, 3, 4, 5 · checked 2026-11-08

Pricing

Published per-token pricing on https://azure.microsoft.com/en-us/pricing/details/azure-openai/ (Standard pay-as-you-go for input and output tokens; Provisioned PTUs with monthly and annual reservations).

Sources 1, 2, 3, 4, 5 · checked 2026-11-08

Published certifications

  • FedRAMP High (Azure OpenAI listed in Azure Government audit scope)
  • DoD IL2, IL4, IL5, IL6 (Azure OpenAI listed in Azure Government audit scope)

Microsoft publishes SOC, ISO/IEC, and HIPAA BAA service scope in the appendices of its Azure Compliance Offerings document, which we did not verify per service. Ask Microsoft for current scope.

Listed as stated by the vendor. Scope can differ by product or subsidiary; request current reports.

Sources 1, 2, 3, 4, 5 · checked 2026-11-08

Limitations to check

  • Abuse monitoring may store flagged prompts and completions for human review unless you are approved for modified abuse monitoring
  • Global deployments may process prompts in any geography where the model is deployed
  • Preview models and features may use different privacy practices

Sources 1, 2, 3, 4, 5 · checked 2026-11-08

Sources

  1. Azure OpenAI pricing: Standard (On-Demand) pay-as-you-go per token; Provisioned PTUs with monthly and annual reservationsazure.microsoft.com/en-us/pricing/details/azure-openai/Checked 2026-11-08
  2. Prompts and completions not available to OpenAI or other providers, not used to train foundation models; Global and DataZone processing locations; modified abuse monitoringlearn.microsoft.com/en-us/azure/foundry/responsible-ai/openai/data-privacyChecked 2026-11-08
  3. v1 API supports the OpenAI client with minimal code changes and chat completions with models from other providerslearn.microsoft.com/en-us/azure/foundry/openai/api-version-lifecycleChecked 2026-11-08
  4. Azure OpenAI listed for FedRAMP High and DoD IL2 to IL6 in Azure Government audit scopelearn.microsoft.com/en-us/azure/azure-government/compliance/azure-services-in-fedramp-auditscopeChecked 2026-11-08
  5. Product page title is Microsoft Foundryazure.microsoft.com/en-us/products/ai-foundryChecked 2026-11-08

Report a correction: [email protected]. Facts are re-checked at least every 90 days.