Microsoft Foundry Updates Bring GPT-5.6 and APAC Zone

Introduction

Microsoft is pushing Microsoft Foundry further into enterprise production scenarios with a set of generally available updates focused on frontier AI models, regional compliance, and agent hosting. For Azure customers building AI agents, this is an important step from pilot projects to scalable, governed deployments.

What’s new in Microsoft Foundry

GPT-5.6 model family now generally available

Microsoft Foundry now includes OpenAI’s GPT-5.6 series:

• GPT-5.6 Sol for advanced reasoning, agentic workflows, and code-heavy tasks
• GPT-5.6 Terra for balanced performance and lower-cost enterprise workloads
• GPT-5.6 Luna for fast, high-volume, latency-sensitive scenarios

These models are available across Standard Global, Data Zones, and other deployment options, giving organizations more flexibility to match cost and performance to specific use cases.

Asia-Pacific Data Zone is now available

Microsoft also announced general availability of the Asia-Pacific Data Zone for Foundry. This gives APAC customers a way to run frontier OpenAI models while keeping data processing within the Asia-Pacific region.

For regulated industries and multinational organizations, this is especially relevant for:

• Data sovereignty requirements
• Regional compliance needs
• Lower-latency AI workloads in APAC

Hosted agents in Foundry Agent Service

Hosted agents in Foundry Agent Service are now generally available, providing a production runtime for agents built with multiple frameworks, including Microsoft Agent Framework, GitHub Copilot SDK, and others.

Key capabilities include:

• VNet integration for network isolation
• Toolboxes for governed tool access
• Voice Live support for real-time voice experiences
• Publishing paths to Microsoft 365 Copilot and Microsoft Teams

Microsoft is also previewing resilient task support for long-running agent workloads, helping agents recover from interruptions without custom recovery logic.

Why this matters for IT administrators

These updates make Foundry more attractive as a single platform for enterprise AI operations. Instead of combining separate tools for models, hosting, governance, and distribution, IT teams can standardize on a Microsoft-managed platform with built-in identity, security, and compliance controls.

For admins, the biggest benefits are clearer governance, regional deployment choices, and easier integration with existing Microsoft 365 and Azure environments.

Next steps

Organizations evaluating enterprise AI agents should:

• Review whether GPT-5.6 Sol, Terra, or Luna best fits their workloads
• Assess whether the APAC Data Zone supports regulatory or residency requirements
• Test Foundry Agent Service for production hosting and Microsoft 365 distribution
• Validate network and security requirements, including VNet integration

Microsoft Foundry is increasingly positioned as Azure’s end-to-end platform for building and operationalizing enterprise AI agents.