Microsoft Foundry Enhances Agentic AI Development with GPT-5.6 Availability, Asia-Pacific Data Zone, and Hosted Agents

Microsoft announced today a significant expansion of its Azure AI Foundry platform, marking a pivotal moment in the widespread adoption and operationalization of AI agents across global enterprises. The platform is now generally available with the integration of OpenAI’s GPT-5.6 model series, the introduction of a dedicated Asia-Pacific (APAC) Data Zone, and the general availability of hosted agents within the Foundry Agent Service. These advancements aim to empower over 100,000 organizations already building on Foundry to move from experimentation to production with greater ease, reliability, and compliance.

The announcement, building upon the foundational promises laid out at Microsoft Build, underscores a strategic vision for the agentic era: enabling developers to build agents within their existing workflows, deploy them on trusted infrastructure, and seamlessly integrate them with users, all without the need for disparate, disconnected platforms. Industry leaders such as Adobe, Telefónica, and Tata Consultancy Services are already leveraging Foundry to run agents in production environments, validating the platform’s capability to deliver tangible business value.

"AI only creates value when it shows up in real systems—systems that are reliable, observable, and aligned to business outcomes," a Microsoft spokesperson stated in a press release accompanying the announcement. "Today, that vision moves from roadmap to reality with three sets of updates now generally available in Microsoft Foundry."

The core of the update lies in bringing frontier models, robust production agent runtime, enterprise-grade identity, security, and compliance controls, and broad distribution across Microsoft 365 into a unified platform. This consolidation is designed to streamline the AI agent lifecycle, from initial development to scaled deployment and ongoing optimization.

Foundry: A Unified Platform for Agentic AI

Microsoft Foundry is positioned as an end-to-end platform specifically engineered for the creation, execution, governance, and distribution of AI agents. The platform is structured around three key pillars:

• Build: Facilitating the development of AI agents using preferred frameworks and models.
• Run: Providing a secure and scalable infrastructure for agent execution.
• Govern: Offering comprehensive controls for managing, optimizing, and ensuring the responsible use of AI agents.

These pillars are further reinforced by the latest Foundry updates, which collectively aim to enable organizations to build, run, and scale production-ready AI agents on a singular, integrated platform.

Empowering Development with Leading Models and Frameworks

The development of AI agents is increasingly integrated into familiar developer environments. Foundry supports this by allowing agents to be built directly within GitHub Copilot and Microsoft Visual Studio (VS) Code. The Foundry Toolkit for VS Code and the Foundry Skill component facilitate the seamless deployment of these agents to the Foundry platform. Whether developers choose to build using the Microsoft Agent Framework, the GitHub Copilot SDK (now generally available), or the Claude Agent SDK, Foundry serves as the definitive production destination.

A critical aspect of agent capability is the underlying reasoning model. Microsoft Foundry provides access to a diverse range of industry-leading models, including frontier, open-source, and task-specific options, through a single platform. This flexibility allows organizations to select the most appropriate model for each specific workload, optimizing for capability, cost, and performance.

GPT-5.6 Series Now Generally Available

A significant highlight of the update is the general availability of OpenAI’s GPT-5.6 model series within Microsoft Foundry Models and the Microsoft Foundry Agent Service. This new series offers three distinct models:

• GPT-5.6 Sol: Designed for complex, high-stakes reasoning and creative tasks, offering the highest level of capability.
• GPT-5.6 Terra: A balanced option providing strong reasoning capabilities with optimized performance for a wide range of business applications.
• GPT-5.6 Luna: Focused on efficiency and speed, ideal for high-throughput, less complex tasks where cost-effectiveness is paramount.

The introduction of the GPT-5.6 series empowers organizations with the granularity to match model capabilities, costs, and performance to specific business scenarios, moving away from a one-size-fits-all approach to AI model selection.

Microsoft emphasized that access to new models is as crucial as their quality. To facilitate widespread adoption, the GPT-5.6 series is available across 28 global regions with Global Standard and Global Priority Processing options. Furthermore, it is supported by Data Zones Standard and Global Provisioned deployments from day one, enabling customers to leverage cutting-edge AI innovations in their existing, deployed, and scaled applications.

GPT-5.6 Pricing Structure (USD per million tokens):

Note: Pricing is indicative and subject to change.

Global Reach and Data Sovereignty with APAC Data Zone

Beyond model availability, the expansion of where these frontier AI models can be run compliantly is equally vital. The general availability of the Asia-Pacific (APAC) Data Zone for Microsoft Foundry addresses this need. This new zone allows customers in the APAC region to process data and run frontier OpenAI models while keeping data processing strictly within regional boundaries. This eliminates the need for separate environments or compromises on capability due to data residency requirements.

With the introduction of Global, Data Zone, and Regional deployment options, Foundry offers organizations the flexibility to align their AI adoption strategies with their specific sovereignty, compliance, performance, and scale mandates, all while maintaining a consistent development and operational experience.

Hongsoo Kim, Chief Data and AI Officer at Viva Republica (Toss), commented on the significance of the APAC Data Zone: "As financial institutions adopt AI, responsible data handling becomes foundational to trust. Microsoft Foundry’s APAC Data Zone allows us to keep data processing regionally anchored while accessing advanced AI models at scale. This gives us the confidence to accelerate AI innovation responsibly and reinforces our ambition to be a leading AI-powered financial platform in Asia."

Driving Impact with Action-Oriented Agents

A powerful model is merely the foundation. To deliver tangible value in production, an AI agent requires a robust runtime environment, an understanding of business context, governed access to tools, persistent memory across interactions, the ability to act on real-world events, and a clear path to end-users. Foundry integrates these capabilities as built-in features, designed to work in concert.

Key components enabling action-oriented agents within Foundry include:

• Hosted Agents: Providing a managed runtime environment that simplifies deployment and scaling.
• Agent Memory: Enabling agents to retain context and learn from past interactions.
• Tooling: Allowing agents to securely access and utilize external services and data sources.
• Eventing: Facilitating agents’ ability to respond to real-time events and triggers.
• Distribution: Offering pathways to deploy agents across Microsoft 365 applications and other enterprise systems.

Governance and Optimization Across the AI Lifecycle

For AI agents to be trusted and effective in production, they must be observable, improvable, and secure. Microsoft Foundry treats trust as a core platform priority, shifting responsibility from individual developers to the platform itself. The latest updates focus on post-build capabilities, enabling teams to monitor agent performance, refine their operations, and demonstrate their value.

Key governance and optimization features include:

• Observability: Providing deep insights into agent behavior, performance, and resource utilization.
• Evaluation: Tools for rigorously testing and validating agent accuracy and effectiveness against defined metrics.
• Cost Management: Features to monitor and control AI spend, ensuring predictable costs as agent usage scales.
• Security and Compliance: Robust controls for data privacy, access management, and adherence to regulatory standards.

As agents scale from pilot programs to handling thousands of requests daily, Foundry equips teams with the tools to maintain predictable expenditure without compromising functionality. This is achieved through a multi-faceted approach:

• Model Choice: Providing flexibility in selecting the most cost-effective and performant model for each task.
• Model Router: Intelligently matches each request to the optimal model based on defined criteria.
• Prompt Caching: Reduces redundant computation by storing and reusing responses to identical prompts.
• PTU Spillover and Quota Optimization: Ensures service continuity during usage spikes and efficient resource allocation.
• Toolboxes: Delivering only the necessary tools for each specific agent request, minimizing overhead.
• Agent Optimizer: Fine-tunes prompts, skills, tools, and model choices against custom evaluators to enhance performance and efficiency.

Beyond cost control, Foundry provides ROI for agents, offering a unified view that connects business value, usage metrics, and operational costs. This enables teams to ascertain whether production agents are generating more value than they consume in resources, and to identify areas where cost might be outpacing return.

A Microsoft Mechanics episode on "Token Economics for Agents" offers a practical walkthrough of these optimization strategies, highlighting the platform’s focus on both performance and financial efficiency.

Real-World Impact: Organizations Building on Foundry

The adoption of Microsoft Foundry extends beyond experimentation, with numerous organizations actively deploying agents to drive business outcomes. These range from digital-native startups to the world’s largest enterprises. Companies like Adobe, Telefónica, and Tata Consultancy Services are cited as examples of entities already running agents in production.

The consistent pattern observed is a significant acceleration in deployment timelines. Teams that previously spent weeks integrating, securing, and deploying AI agents can now achieve the same in days. This rapid deployment is facilitated by an infrastructure that meets stringent compliance requirements and allows agents to reach users through familiar tools.

Getting Started with Microsoft Foundry

All the newly announced capabilities are live and accessible within Microsoft Foundry. Developers can begin their journey by consulting the comprehensive documentation and Microsoft Learn courses available. A quickstart guide provides a step-by-step walkthrough for setting up, testing, and deploying a production-ready hosted agent.

For those seeking a structured learning path, resources such as "AI Agents for Beginners" (a 12-lesson curriculum), guided labs like "Develop AI Agents in Azure," "Hosted Agents Workshop (.NET)," and "Foundry Toolkit for VS Code and hosted agents workshop," and the "ZavaShop Supply Chain Workshop" are available. Furthermore, a practical guide on "Evaluating AI Agents with Microsoft Foundry" offers insights into ensuring agent quality.

A detailed explanation of Foundry Agent Service and the Microsoft Agent Framework is available in a Microsoft Mechanics video, featuring Jeff Hollan, who walks through operationalizing AI agents from deployment to real-world impact. This wealth of resources aims to lower the barrier to entry and accelerate the adoption of advanced AI agent technologies for businesses worldwide.