Azure OpenAI models and regions for Foundry Agent Service (classic)

Azure OpenAI models power agents in Foundry Agent Service. To use these models, you need a Microsoft Foundry project with access to Agent Service. Use the tabs to find a supported model, deployment type, and region combination. For details on deployment types, see Deployment types for Microsoft Foundry Models.

Agents (classic) are deprecated. To use models later than gpt-5, see the agents (new) documentation.

Available models

• gpt-5 family (gpt-5, gpt-5-mini, gpt-5-nano, gpt-5-chat): Frontier-scale reasoning for complex, multi-step tasks. Registration is required. These models can use only the code interpreter and file search tools.
• gpt-4.1 family (gpt-4.1, gpt-4.1-mini, gpt-4.1-nano): Cost-effective models for general-purpose agent workloads.
• gpt-4o family (gpt-4o, gpt-4o-mini): Multimodal capabilities with vision support.
• gpt-4 and gpt-35-turbo: Legacy models for backward compatibility.

Non-OpenAI models

In addition to Azure OpenAI models, you can use models sold directly by Azure. These models offer specialized capabilities for specific use cases, such as deterministic reasoning or high-throughput generation.

Models sold directly by Azure:

• MAI-DS-R1: Deterministic, precision-focused reasoning.
• grok-4: Frontier-scale reasoning for complex, multiple-step problem solving.
• grok-4-fast-reasoning: Accelerated agentic reasoning optimized for workflow automation.
• grok-4-fast-non-reasoning: High-throughput, low-latency generation and system routing.
• grok-3: Strong reasoning for complex, system-level workflows.
• grok-3-mini: Lightweight model optimized for interactive, high-volume use cases.
• Llama-3.3-70B-Instruct: Versatile model for enterprise Q&A, decision support, and system orchestration.
• Llama-4-Maverick-17B-128E-Instruct-FP8: FP8-optimized model that delivers fast, cost-efficient inference.
• DeepSeek-V3-0324: Multimodal understanding across text and images.
• DeepSeek-V3.1: Enhanced multimodal reasoning and grounded retrieval.
• DeepSeek-R1-0528: Advanced long-form and multiple-step reasoning.
• gpt-oss-120b: Open-ecosystem model that supports transparency and reproducibility.

Verify model support

Model availability can change over time. To check what you can deploy for your project and region:

1. Sign in to Microsoft Foundry. Make sure the New Foundry toggle is off. These steps refer to Foundry (classic).
2. Go to the Model catalog.
3. Filter the models by Capabilities and select Agent supported.

If you use provisioned throughput, make sure you have provisioned throughput units (PTUs) available in the target region. For background, see Provisioned throughput.

Troubleshooting

A model or version isn't available in your region

• Confirm you selected the right tab for your deployment type.
• Try a different region that supports the model and version.
• If you're using gpt-5 models, make sure your subscription has access. Some models require registration.

File search isn't available

• File search isn't available in Italy North and Brazil South. Choose a supported region, or use a different tool.

Provisioned throughput deployment fails

• Confirm you have enough PTUs available in the region.
• Review Provisioned throughput and Spillover traffic management.

Related content

• Create a new agent
• Foundry Agent Service quotas and limits
• Deployment types for Microsoft Foundry Models
• Feature availability across cloud regions