From 56e8b59cb401b32fa5fa8d049bb1be99a966c9a3 Mon Sep 17 00:00:00 2001 From: "github-actions[bot]" Date: Tue, 29 Sep 2026 06:06:54 +0000 Subject: [PATCH] docs: sync and convert from upstream Automated daily sync from MicrosoftDocs/azure-ai-docs-pr Co-authored-by: github-actions[bot] --- .../privacy-compliance-faq.mdx | 2 +- docs/models/claude-models-quotas-limits.mdx | 12 ++++++------ docs/models/claude-models.mdx | 8 ++++---- docs/models/deploy-models-managed.mdx | 4 +++- docs/models/deployments-overview.mdx | 3 +-- docs/models/hugging-face-models.mdx | 14 +++++++------- docs/models/managed-compute-overview.mdx | 2 +- docs/models/models-from-partners.mdx | 8 ++++---- 8 files changed, 27 insertions(+), 26 deletions(-) diff --git a/docs/developer-tools-and-integrations/privacy-compliance-faq.mdx b/docs/developer-tools-and-integrations/privacy-compliance-faq.mdx index e17bf437..3a9176c0 100644 --- a/docs/developer-tools-and-integrations/privacy-compliance-faq.mdx +++ b/docs/developer-tools-and-integrations/privacy-compliance-faq.mdx @@ -25,7 +25,7 @@ No, customers don't require a separate BAA with Fireworks when using the Firewor ## Does customer data travel outside of Azure when using the Fireworks on Foundry service? Why? What data processing terms apply? -Yes, customer data travels outside Microsoft facilities because Fireworks runs inferencing for this service on Fireworks’ GPUs. However, since Fireworks is a Microsoft subprocessor, any processing of customer data continues to be governed by Microsoft’s Product Terms and DPA, and any documented exceptions (see response to #1 above). +Yes, customer data travels outside Microsoft facilities because Fireworks runs inferencing for this service on Fireworks’ GPUs. However, since Fireworks is a Microsoft subprocessor, any processing of customer data continues to be governed by Microsoft’s Product Terms and DPA, and any documented exceptions. ## Where is the customer data stored? Do Foundry’s DataZone commitments apply? diff --git a/docs/models/claude-models-quotas-limits.mdx b/docs/models/claude-models-quotas-limits.mdx index 5dc8e917..70b0264d 100644 --- a/docs/models/claude-models-quotas-limits.mdx +++ b/docs/models/claude-models-quotas-limits.mdx @@ -59,8 +59,8 @@ Your Azure subscription type determines your default rate limits. The **Version | claude-opus-4-7 | Global Standard | N/A | Yes | 40 | 40,000 | 8,000 | | claude-opus-4-6 | Global Standard | N/A | Yes | 40 | 40,000 | 8,000 | | claude-opus-4-5 | Global Standard | N/A | Yes | 40 | 40,000 | 8,000 | - | claude-sonnet-5.5 | Global Standard | Yes | Yes | 40 | 40,000 | 8,000 | - | claude-sonnet-5.5 | Data Zone Standard (US) | Yes | N/A | 40 | 40,000 | 8,000 | + | claude-sonnet-5-5 | Global Standard | Yes | Yes | 40 | 40,000 | 8,000 | + | claude-sonnet-5-5 | Data Zone Standard (US) | Yes | N/A | 40 | 40,000 | 8,000 | | claude-sonnet-5 | Global Standard | Yes | Yes | 40 | 40,000 | 8,000 | | claude-sonnet-5 | Data Zone Standard (US) | Yes | N/A | 40 | 40,000 | 8,000 | | claude-sonnet-4-6 | Global Standard | N/A | Yes | 80 | 80,000 | 16,000 | @@ -85,8 +85,8 @@ Your Azure subscription type determines your default rate limits. The **Version | claude-opus-4-7 | Global Standard | N/A | Yes | 10,000 | 10,000,000 | 2,000,000 | | claude-opus-4-6 | Global Standard | N/A | Yes | 10,000 | 10,000,000 | 2,000,000 | | claude-opus-4-5 | Global Standard | N/A | Yes | 10,000 | 10,000,000 | 2,000,000 | - | claude-sonnet-5.5 | Global Standard | Yes | Yes | 10,000 | 10,000,000 | 2,000,000 | - | claude-sonnet-5.5 | Data Zone Standard (US) | Yes | N/A | 10,000 | 10,000,000 | 2,000,000 | + | claude-sonnet-5-5 | Global Standard | Yes | Yes | 10,000 | 10,000,000 | 2,000,000 | + | claude-sonnet-5-5 | Data Zone Standard (US) | Yes | N/A | 10,000 | 10,000,000 | 2,000,000 | | claude-sonnet-5 | Global Standard | Yes | Yes | 10,000 | 10,000,000 | 2,000,000 | | claude-sonnet-5 | Data Zone Standard (US) | Yes | N/A | 10,000 | 10,000,000 | 2,000,000 | | claude-sonnet-4-6 | Global Standard | N/A | Yes | 10,000 | 10,000,000 | 2,000,000 | @@ -111,8 +111,8 @@ Your Azure subscription type determines your default rate limits. The **Version | claude-opus-4-7 | Global Standard | N/A | Yes | 0 | 0 | 0 | | claude-opus-4-6 | Global Standard | N/A | Yes | 0 | 0 | 0 | | claude-opus-4-5 | Global Standard | N/A | Yes | 0 | 0 | 0 | - | claude-sonnet-5.5 | Global Standard | Yes | Yes | 0 | 0 | 0 | - | claude-sonnet-5.5 | Data Zone Standard (US) | Yes | N/A | 0 | 0 | 0 | + | claude-sonnet-5-5 | Global Standard | Yes | Yes | 0 | 0 | 0 | + | claude-sonnet-5-5 | Data Zone Standard (US) | Yes | N/A | 0 | 0 | 0 | | claude-sonnet-5 | Global Standard | Yes | Yes | 0 | 0 | 0 | | claude-sonnet-5 | Data Zone Standard (US) | Yes | N/A | 0 | 0 | 0 | | claude-sonnet-4-6 | Global Standard | N/A | Yes | 0 | 0 | 0 | diff --git a/docs/models/claude-models.mdx b/docs/models/claude-models.mdx index 6e1c9a4d..2c9a1218 100644 --- a/docs/models/claude-models.mdx +++ b/docs/models/claude-models.mdx @@ -61,7 +61,7 @@ For errors you might encounter when you deploy or call Claude models, see [Deplo | `claude-opus-4-7` | - Hosted on Anthropic infrastructure: GA | 1M / 128K | - Adaptive thinking
- Reasoning over entire codebases- High-resolution image input (up to 2576px / 3.75MP)
- See [Migrating to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide) | - Coding
- Enterprise workflows
- Long-running agents
- Multimodal reasoning
- Financial analysis
- Cybersecurity | | `claude-opus-4-6` | - Hosted on Anthropic infrastructure: GA | 1M / 128K | - Adaptive thinking
- Image and text input
- Computer use
- Advanced tool use (search, programmatic calling, examples)
- See [Migrating to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide) | - Coding
- Enterprise agents | | `claude-opus-4-5` | - Hosted on Anthropic infrastructure: GA | 200K / 64K | - Extended thinking
- Image and text input
- Computer use
- Advanced tool use (search, programmatic calling, examples)
- See [Migrating to Claude Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide) | - Coding
- Agents
- Computer use
- Enterprise workflows | -| `claude-sonnet-5.5` | - Hosted on Azure: GA
- Hosted on Anthropic infrastructure: GA | 1M / 128K | - Adaptive thinking
- `xhigh` effort level
- Reasoning over entire codebases and multi-day project context
- High-res image input (up to 2576px / 3.75MP) are on by default
- Mid-conversation3 `role:"system"`
- Token budgets3 (`task_budget`) | - Coding
- Long-running agents
- Financial analysis
- Cybersecurity
- Computer use | +| `claude-sonnet-5-5` | - Hosted on Azure: GA
- Hosted on Anthropic infrastructure: GA | 1M / 128K | - Adaptive thinking
- `xhigh` effort level
- Reasoning over entire codebases and multi-day project context
- High-res image input (up to 2576px / 3.75MP) are on by default
- Mid-conversation3 `role:"system"`
- Token budgets3 (`task_budget`) | - Coding
- Long-running agents
- Financial analysis
- Cybersecurity
- Computer use
- See [What's new in Claude Sonnet 5.5](https://platform.claude.com/docs/en/models/sonnet-5-5/whats-new-sonnet-5-5) | | `claude-sonnet-5` | - Hosted on Azure: GA
- Hosted on Anthropic infrastructure: GA | 1M / 128K | - Adaptive thinking
- `xhigh` effort level
- Reasoning over entire codebases and multi-day project context
- High-res image input (up to 2576px / 3.75MP) are on by default
- Mid-conversation3 `role:"system"`
- Token budgets3 (`task_budget`)
- See [What's new in Claude Sonnet 5](https://platform.claude.com/docs/en/about-claude/models/whats-new-sonnet-5) | - Coding
- Long-running agents
- Financial analysis
- Cybersecurity
- Computer use | | `claude-sonnet-4-6` | - Hosted on Anthropic infrastructure: GA | 1M / 128K | - Adaptive thinking
- Image and text input
- Computer use
- Advanced tool use (search, programmatic calling, examples)
- See [Migrating to Claude Sonnet 5](https://platform.claude.com/docs/en/about-claude/models/migration-guide#migrating-to-claude-sonnet-5) | - Coding
- Agents
- Enterprise workflows | | `claude-sonnet-4-5` | - Hosted on Anthropic infrastructure: GA | 200K / 64K | - Extended thinking
- Image and text input
- Computer use
- See [Migrating to Claude Sonnet 5](https://platform.claude.com/docs/en/about-claude/models/migration-guide#migrating-to-claude-sonnet-5) | - Agents and complex, long-horizon tasks
- High-volume workloads | @@ -144,7 +144,7 @@ The **Thinking** feature allows specific values for the `thinking` parameter typ | `claude-opus-4-8` | Yes | No | Yes | | `claude-opus-4-7` | Yes | No | Yes | | `claude-opus-4-6` | Yes | Yes | Yes | -| `claude-sonnet-5.5` | Yes | Yes | Yes | +| `claude-sonnet-5-5` | Yes | Yes | Yes | | `claude-sonnet-5` | Yes | No | Yes | | `claude-sonnet-4-6` | Yes | Yes | Yes | @@ -163,7 +163,7 @@ The **Effort** feature allows specific `effort` levels for each model, as descri | `claude-opus-4-8` | Yes | Yes | Yes | Yes | Yes | | `claude-opus-4-7` | Yes | Yes | Yes | Yes | Yes | | `claude-opus-4-6` | Yes | Yes | Yes | No | Yes | -| `claude-sonnet-5.5` | Yes | Yes | Yes | Yes | Yes | +| `claude-sonnet-5-5` | Yes | Yes | Yes | Yes | Yes | | `claude-sonnet-5` | Yes | Yes | Yes | Yes | Yes | | `claude-sonnet-4-6` | Yes | Yes | Yes | No | Yes | @@ -217,7 +217,7 @@ Manage the documents and data you provide to Claude. Claude models in Foundry are available for the following deployment types in specific Azure regions: - **Global Standard**: All Claude models (Hosted on Azure and Hosted on Anthropic infrastructure). -- **Data Zone Standard (US)**: Hosted on Azure versions of `claude-opus-5-5`, `claude-opus-5`, `claude-opus-4-8`, `claude-sonnet-5`, and `claude-sonnet-5.5`. +- **Data Zone Standard (US)**: Hosted on Azure versions of `claude-opus-5-5`, `claude-opus-5`, `claude-opus-4-8`, `claude-sonnet-5`, and `claude-sonnet-5-5`. For the exact Azure regions where Claude models are available for deployment, see [Region availability by deployment type](/models/models-from-partners#region-availability-by-deployment-type). diff --git a/docs/models/deploy-models-managed.mdx b/docs/models/deploy-models-managed.mdx index 81d45ca1..e91ff52f 100644 --- a/docs/models/deploy-models-managed.mdx +++ b/docs/models/deploy-models-managed.mdx @@ -56,6 +56,8 @@ Managed compute in Foundry is in **public preview**. APIs, SKU names, and suppor Managed compute deploys models from the **Hugging Face Collection** in the Foundry model catalog, served from the `azure-huggingface` registry. +For Hugging Face model availability in your project and Hugging Face-specific deployment guidance, see [Hugging Face models in Microsoft Foundry](/models/hugging-face-models). + 1. Sign in to [Microsoft Foundry](https://ai.azure.com/?cid=learnDocs). Make sure the **New Foundry** toggle is on. These steps refer to **Foundry (new)**. @@ -86,7 +88,7 @@ A model ID and a deployment template ID must be compatible; every template lists To learn more about deployment templates, see [Deployment template](/models/managed-compute-overview#deployment-template) in the Managed compute overview article. -- **Accelerator type**: for example `H100_80GB`, `A100_80GB`, or `MI_300_192GB`. Shown next to each template in the deployment wizard. +- **Accelerator type**: for example, `H100_80GB`, `A100_80GB`, or `MI300_192GB`. Shown next to each template in the deployment wizard. ## Deploy the model diff --git a/docs/models/deployments-overview.mdx b/docs/models/deployments-overview.mdx index 9be4edc7..ff98c75f 100644 --- a/docs/models/deployments-overview.mdx +++ b/docs/models/deployments-overview.mdx @@ -92,13 +92,12 @@ Managed compute (Preview) supports: - **Model-instance sizing** — Deployments are sized in model-centric terms. You don't need to pick virtual machine SKUs, because Foundry chooses GPUs per instance based on model size, architecture, context length, and whether the workload is optimized for latency or throughput. - **Optimized inference runtimes** — Microsoft-curated vLLM, SGLang, and NVIDIA NIM containers with continuous batching and tensor parallelism. - **Accelerator families** — A100 (80 GB), H100 (80 GB), and MI300X (192 GB). -- **Auto-scaling and scale-to-zero** — Auto-scale from live traffic or scale manually. Configure an idle timeout so the deployment scales to zero when no traffic arrives, making billing stop immediately. - **Microsoft-managed runtimes** — Microsoft owns serving runtimes, base container images, and security patches. Updates are applied to live deployments automatically. - **Observability metrics** — Each deployment emits API call count by status code and response-time percentiles. Chat-completion models also emit input and output token counts, time-to-first-token (TTFT) percentiles, and total response-time percentiles, grouped by time. ### Billing and quota -Managed compute billing is hourly per accelerator SKU, with throughput per GPU as the underlying billing unit. Auto-scale and scale-to-zero align cost with actual traffic so that billing stops immediately when instances scale down. +Managed compute billing is hourly per accelerator SKU, with throughput per GPU as the underlying billing unit. Quota is granted per accelerator SKU per region through the **Foundry quota process** and is **separate from Azure VM quota**. Azure virtual machines are an infrastructure-as-a-service (IaaS) offering with regional SKUs; managed compute is a PaaS offering that leads with Global and Data Zone processing. Existing Azure VM quota can't be applied to a managed compute deployment. diff --git a/docs/models/hugging-face-models.mdx b/docs/models/hugging-face-models.mdx index 09bdac24..8fb2c1a3 100644 --- a/docs/models/hugging-face-models.mdx +++ b/docs/models/hugging-face-models.mdx @@ -1,9 +1,9 @@ --- -title: "Hugging Face models in Microsoft Foundry (preview)" -description: "Learn how to discover and deploy Hugging Face models in Microsoft Foundry (preview), including project availability filtering, managed compute behavior, and classic fallback for unavailable models." +title: "Hugging Face models in Microsoft Foundry" +description: "Learn how to discover and deploy Hugging Face models in Microsoft Foundry with Foundry Managed Compute (preview), including project availability filtering and classic fallback for unavailable models." --- -Hugging Face-published models are available in the Foundry model catalog for deployment in Microsoft Foundry. In the current Foundry experience, these models use managed compute. +Hugging Face-published models are available in the Foundry model catalog for deployment in Microsoft Foundry. In the current Foundry experience, these models use Foundry Managed Compute (preview). In this article, you discover deployable Hugging Face models, deploy one with managed compute, and call the endpoint from your application. @@ -11,7 +11,7 @@ In model discovery, use the **Available in my Project** filter to find the Huggi -Foundry managed compute is currently in preview. Preview features might not be available in all regions and are subject to supplemental terms. For more information, see [Supplemental Terms of Use for Microsoft Azure Previews](https://azure.microsoft.com/support/legal/preview-supplemental-terms/). +This preview feature might not be available in all regions and is subject to supplemental terms. For more information, see [Supplemental Terms of Use for Microsoft Azure Previews](https://azure.microsoft.com/support/legal/preview-supplemental-terms/). Hugging Face models in Foundry are globally available. For general service availability information by region, see [Azure products by region](https://azure.microsoft.com/explore/global-infrastructure/products-by-region/). @@ -34,7 +34,7 @@ Models from Hugging Face are subject to third-party license terms available on t - **Cognitive Services Contributor** (or **Foundry Owner** / **Foundry Account Owner**) to create, update, and delete managed compute deployments. - **Foundry User** to call the deployment with Microsoft Entra ID from the playground, SDK, or REST. For role definitions, see [Role-based access control in Foundry](../../concepts/rbac-foundry#managed-compute-control-plane-operations). -- Available GPU managed compute quota in your Azure subscription for the selected accelerator family. Foundry managed compute quota is separate from Azure VM quota and uses a different request path. In the [Foundry portal](https://ai.azure.com/nextgen), go to **Manage** > **Quota** > **Managed compute** > **Request quota** to check your current allocation or request an increase. For detailed guidance, see [Request more quota](/models/deploy-models-managed#request-more-quota). +- Available GPU quota for Foundry Managed Compute (preview) in your Azure subscription for the selected accelerator family. This quota is separate from Azure VM quota and uses a different request path. In the [Foundry portal](https://ai.azure.com/nextgen), go to **Manage** > **Quota** > **Managed compute** > **Request quota** to check your current allocation or request an increase. For detailed guidance, see [Request more quota](/models/deploy-models-managed#request-more-quota). ## Deploy a Hugging Face model @@ -122,7 +122,7 @@ After the deployment finishes, validate the endpoint before you integrate it int 1. Run a test inference from the deployment details page to confirm that the model returns a successful response for your task type. -1. If deployment fails, review the deployment logs. Then, verify GPU quota and regional availability for the selected model and accelerator type. Foundry managed compute quota is separate from Azure VM quota. For quota checks and increase requests, see [Request more quota](/models/deploy-models-managed#request-more-quota). +1. If deployment fails, verify GPU quota. Quota for Foundry Managed Compute (preview) is separate from Azure VM quota. For quota checks and increase requests, see [Request more quota](/models/deploy-models-managed#request-more-quota). ## Invoke the endpoint from your application @@ -216,7 +216,7 @@ At a high level: - You discover Hugging Face models in the model catalog. - You filter to models that are deployable in your current project. - You deploy a supported model; Foundry provisions dedicated GPU compute instances and exposes an endpoint. -- You invoke the deployment through the endpoint by using Foundry managed compute pricing. +- You invoke the deployment through the endpoint. Managed compute provides dedicated GPU compute instances, endpoint-based inference access, and billing through the Foundry Models Managed Compute pricing model. For current rates and billing details, see [Foundry Models pricing - Managed Compute](https://azure.microsoft.com/pricing/details/ai-foundry-models/microsoft/#pricing). diff --git a/docs/models/managed-compute-overview.mdx b/docs/models/managed-compute-overview.mdx index 30b36f5a..579376ee 100644 --- a/docs/models/managed-compute-overview.mdx +++ b/docs/models/managed-compute-overview.mdx @@ -62,7 +62,7 @@ Managed compute deployments target an **accelerator family**, not a specific vir - NVIDIA A100 80 GB (`A100_80GB`) - NVIDIA H100 80 GB (`H100_80GB`) -- AMD MI300X 192 GB (`MI_300_192GB`) +- AMD MI300X 192 GB (`MI300_192GB`) Quota is granted per accelerator family per region. diff --git a/docs/models/models-from-partners.mdx b/docs/models/models-from-partners.mdx index 5c3cf9bd..c4fe32d2 100644 --- a/docs/models/models-from-partners.mdx +++ b/docs/models/models-from-partners.mdx @@ -92,7 +92,7 @@ For a list of supported regions, see [supported geographic locations](/partner-c | `claude-opus-4-7` | Messages | - **Input:** text, image, and code
- **Output:** text, image, and code (128,000 max tokens)
- **Context window:** 1,000,000
- **Languages:** `en`, `fr`, `ar`, `zh`, `ja`, `ko`, `es`, `hi`
- **Tool calling:** Yes (file search and code execution)
- **Response formats:** Text in various formats (e.g., prose, lists, Markdown tables, JSON, HTML, code in various programming languages)
- **Key parameters:**
`top_k`, `temperature`, and `thinking={"type":"enabled"}` are **not supported**.
`top_p` must be 0.99. When omitted, the default (0.99) is used. | [Microsoft Managed Countries/Regions](/partner-center/marketplace/tax-details-marketplace#microsoft-managed-countriesregions) (except Belarus and Russia) | | `claude-opus-4-6` | Messages | - **Input:** text, image, and code
- **Output:** text, image, and code (128,000 max tokens)
- **Context window:** 1,000,000
- **Languages:** `en`, `fr`, `ar`, `zh`, `ja`, `ko`, `es`, `hi`
- **Tool calling:** Yes (file search and code execution)
- **Response formats:** Text in various formats (e.g., prose, lists, Markdown tables, JSON, HTML, code in various programming languages) | [Microsoft Managed Countries/Regions](/partner-center/marketplace/tax-details-marketplace#microsoft-managed-countriesregions) (except Belarus and Russia) | | `claude-opus-4-5` | Messages | - **Input:** text, image, and code
- **Output:** text (64,000 max tokens)
- **Context window:** 200,000
- **Languages:** `en`, `fr`, `ar`, `zh`, `ja`, `ko`, `es`, `hi`
- **Tool calling:** Yes (file search and code execution)
- **Response formats:** Text in various formats (e.g., prose, lists, Markdown tables, JSON, HTML, code in various programming languages) | [Microsoft Managed Countries/Regions](/partner-center/marketplace/tax-details-marketplace#microsoft-managed-countriesregions) (except Belarus and Russia) | -| `claude-sonnet-5.5` | Messages | - **Input:** text, image, and code
- **Output:** text, image, and code (128,000 max tokens)
- **Context window:** 1,000,000
- **Languages:** `en`, `fr`, `ar`, `zh`, `ja`, `ko`, `es`, `hi`
- **Tool calling:** Yes (file search and code execution)
- **Response formats:** Text in various formats (e.g., prose, lists, Markdown tables, JSON, HTML, code in various programming languages)
- **Key parameters:**
`top_k`, `temperature`, and `thinking={"type":"enabled"}` are **not supported**.
`output_format` supported only for `thinking={"type":"adaptive"}`.
`top_p` must be 0.99. When omitted, the default (0.99) is used. | [Microsoft Managed Countries/Regions](/partner-center/marketplace/tax-details-marketplace#microsoft-managed-countriesregions) (except Belarus and Russia)
US (Hosted on Azure for Data Zone Standard) | +| `claude-sonnet-5-5` | Messages | - **Input:** text, image, and code
- **Output:** text, image, and code (128,000 max tokens)
- **Context window:** 1,000,000
- **Languages:** `en`, `fr`, `ar`, `zh`, `ja`, `ko`, `es`, `hi`
- **Tool calling:** Yes (file search and code execution)
- **Response formats:** Text in various formats (e.g., prose, lists, Markdown tables, JSON, HTML, code in various programming languages)
- **Key parameters:**
`top_k`, `temperature`, and `thinking={"type":"enabled"}` are **not supported**.
`output_format` supported only for `thinking={"type":"adaptive"}`.
`top_p` must be 0.99. When omitted, the default (0.99) is used. | [Microsoft Managed Countries/Regions](/partner-center/marketplace/tax-details-marketplace#microsoft-managed-countriesregions) (except Belarus and Russia)
US (Hosted on Azure for Data Zone Standard) | | `claude-sonnet-5` | Messages | - **Input:** text, image, and code
- **Output:** text, image, and code (128,000 max tokens)
- **Context window:** 1,000,000
- **Languages:** `en`, `fr`, `ar`, `zh`, `ja`, `ko`, `es`, `hi`
- **Tool calling:** Yes (file search and code execution)
- **Response formats:** Text in various formats (e.g., prose, lists, Markdown tables, JSON, HTML, code in various programming languages)
- **Key parameters:**
`top_k`, `temperature`, and `thinking={"type":"enabled"}` are **not supported**.
`output_format` supported only for `thinking={"type":"adaptive"}`.
`top_p` must be 0.99. When omitted, the default (0.99) is used. | [Microsoft Managed Countries/Regions](/partner-center/marketplace/tax-details-marketplace#microsoft-managed-countriesregions) (except Belarus and Russia)
US (Hosted on Azure for Data Zone Standard) | | `claude-sonnet-4-6` | Messages | - **Input:** text, image, and code
- **Output:** text, image, and code (128,000 max tokens)
- **Context window:** 1,000,000
- **Languages:** `en`, `fr`, `ar`, `zh`, `ja`, `ko`, `es`, `hi`
- **Tool calling:** Yes (file search and code execution)
- **Response formats:** Text in various formats (e.g., prose, lists, Markdown tables, JSON, HTML, code in various programming languages) | [Microsoft Managed Countries/Regions](/partner-center/marketplace/tax-details-marketplace#microsoft-managed-countriesregions) (except Belarus and Russia) | | `claude-sonnet-4-5` | Messages | - **Input:** text, image, and code
- **Output:** text (64,000 max tokens)
- **Context window:** 200,000
- **Languages:** `en`, `fr`, `ar`, `zh`, `ja`, `ko`, `es`, `hi`
- **Tool calling:** Yes (file search and code execution)
- **Response formats:** Text in various formats (e.g., prose, lists, Markdown tables, JSON, HTML, code in various programming languages) | [Microsoft Managed Countries/Regions](/partner-center/marketplace/tax-details-marketplace#microsoft-managed-countriesregions) (except Belarus and Russia) | @@ -204,7 +204,7 @@ For **global deployments**, Microsoft processes prompts and responses in any Azu | claude-sonnet-4-6 | 1 | - | - | - | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | - | ✅ | | claude-sonnet-5 | 1 | - | - | - | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | - | ✅ | | claude-sonnet-5 | 2 | - | - | - | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | - | ✅ | - | claude-sonnet-5.5 | 1 | - | - | - | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | - | ✅ | + | claude-sonnet-5-5 | 1 | - | - | - | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | - | ✅ | | Codestral-2501 | 2 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | | Llama-4-Scout-17B-16E-Instruct | 1 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | | Ministral-3B | 1 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | @@ -234,7 +234,7 @@ For **global deployments**, Microsoft processes prompts and responses in any Azu | claude-sonnet-4-6 | 1 | - | - | - | - | - | - | ✅ | - | - | - | - | - | | claude-sonnet-5 | 1 | - | - | - | - | - | - | ✅ | - | - | - | - | - | | claude-sonnet-5 | 2 | - | - | - | - | - | - | ✅ | - | - | - | - | - | - | claude-sonnet-5.5 | 1 | - | - | - | - | - | - | ✅ | - | - | - | - | - | + | claude-sonnet-5-5 | 1 | - | - | - | - | - | - | ✅ | - | - | - | - | - | | Codestral-2501 | 2 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | | Llama-4-Scout-17B-16E-Instruct | 1 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | | Ministral-3B | 1 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | @@ -282,7 +282,7 @@ For **Data Zone** deployments, Microsoft processes prompts and responses anywher | claude-opus-5 | 2 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | | claude-opus-5-5 | 2 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | | claude-sonnet-5 | 2 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | - | claude-sonnet-5.5 | 2 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | + | claude-sonnet-5-5 | 2 | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |