Microsoft Foundry is now generally available, featuring GPT-5.6, the Asia-Pacific Data Zone, and hosted agents in Foundry Agent Service, marking a significant leap forward in enterprise-grade AI agent development and deployment. This advancement aims to bridge the gap between experimental AI and tangible business value by integrating powerful models, robust infrastructure, and seamless distribution channels into a unified platform. Over 100,000 organizations are already leveraging Microsoft Foundry, with prominent companies such as Adobe, Telefónica, and Tata Consultancy Services actively deploying agents in production environments.
The announcement, building on promises made at Microsoft Build, underscores a commitment to enabling developers to build agents within their existing workflows, deploy them on trusted infrastructure, and deliver them to end-users without the need for disparate, disconnected platforms. This transition from roadmap to reality signifies a pivotal moment for organizations seeking to harness the power of AI agents for real-world applications. The newly available capabilities encompass frontier models, a production-ready agent runtime, enterprise-grade identity and security controls, compliance assurances, and integration across Microsoft 365, all within a singular, cohesive platform.
Foundry: The Cornerstone of Enterprise AI Agent Development
Microsoft Foundry is positioned as an end-to-end platform designed to facilitate the building, running, governing, and distribution of AI agents. Its architecture is built upon three core pillars, enabling organizations to navigate the entire lifecycle of AI agent deployment:
- Build: Providing developers with familiar tools and frameworks to create sophisticated AI agents.
- Run: Offering a robust and scalable infrastructure for agent execution.
- Govern & Distribute: Ensuring security, compliance, and efficient delivery of agents to end-users.
These pillars converge in the latest Foundry updates, empowering organizations to move from initial development to large-scale production with unprecedented ease and efficiency.
Empowering Developers: Seamless Integration and Model Flexibility
Agent development is streamlined by integrating directly with widely adopted developer tools such as GitHub Copilot and Microsoft Visual Studio (VS) Code. The Foundry Toolkit for VS Code and the Foundry skill simplify the deployment process, allowing teams to work within their preferred environments. Foundry serves as the ultimate production destination, supporting various development frameworks including Microsoft Agent Framework, GitHub Copilot SDK, and Claude Agent SDK.
The efficacy of an AI agent is intrinsically linked to the underlying model’s reasoning capabilities. Microsoft Foundry addresses this by providing unified access to a diverse array of industry-leading frontier, open-source, and task-specific models. This flexibility allows organizations to select the optimal model for each unique workload, moving beyond a one-size-fits-all approach.
A significant announcement today is the general availability of OpenAI’s GPT-5.6 series within Microsoft Foundry Models and Microsoft Foundry Agent Service. This series includes:
- GPT-5.6 Sol: The most advanced iteration, offering unparalleled reasoning and contextual understanding for complex tasks.
- GPT-5.6 Terra: A balanced model providing high performance and broad applicability across a range of business scenarios.
- GPT-5.6 Luna: Optimized for efficiency and cost-effectiveness, suitable for high-volume, less complex tasks.
This tiered approach empowers organizations to meticulously match model capabilities, costs, and performance metrics to specific business needs, fostering a more strategic and economical AI implementation. The availability of GPT-5.6 across 28 global regions, including Standard, Priority, and Data Zone deployments, ensures that customers can leverage cutting-edge AI innovations precisely where their applications are already deployed and scaled.
The pricing structure for GPT-5.6 reflects this tiered approach, offering different cost points for input and output tokens across its various models and deployment options:
| Model | Deployment | Pricing (USD $/million tokens) |
|---|---|---|
| GPT-5.6 Sol | Standard Global | Input: 5.00, Output: 30.00 |
| GPT-5.6 Terra | Standard Global | Input: 2.50, Output: 15.00 |
| GPT-5.6 Luna | Standard Global | Input: 1.00, Output: 6.00 |
Global Reach and Data Sovereignty: The Asia-Pacific Data Zone
The expansion of AI capabilities is not solely about the availability of more models, but also about ensuring their compliant execution in diverse geographical locations. The newly announced general availability of the Asia-Pacific (APAC) Data Zone for Microsoft Foundry addresses this critical requirement. This initiative allows organizations in the APAC region to utilize advanced OpenAI models while ensuring that data processing remains within regional boundaries. This eliminates the need for complex, separate environments and ensures that regional data sovereignty and compliance mandates are met without compromising on the cutting-edge capabilities of frontier AI.
Foundry’s comprehensive deployment options—Global, Data Zone, and Regional—provide organizations with the granular control needed to align their AI adoption strategies with their specific sovereignty, compliance, performance, and scalability requirements, all while maintaining a consistent development and operational experience across all environments.
Hongsoo Kim, Chief Data and AI Officer at Viva Republica (Toss), highlighted the significance of these advancements: "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
Beyond possessing a powerful model, deploying an AI agent effectively in a production environment requires a robust runtime, business-specific knowledge, governed access to tools, persistent memory across interactions, and a clear pathway to end-users. Foundry integrates these essential components as built-in capabilities, designed to work harmoniously:
- Production Runtime: A stable and scalable environment for agents to operate.
- Business Knowledge Integration: Enabling agents to access and utilize organizational data.
- Tool Orchestration: Secure and efficient management of agent interactions with external tools and services.
- Conversational Memory: Allowing agents to retain context across multiple interactions.
- Event-Driven Actions: Enabling agents to respond to real-world events and triggers.
- Distribution Channels: Seamless integration with platforms like Microsoft 365 for widespread user access.
Governance and Optimization: Ensuring Trust and Efficiency
The ability to observe, improve, and secure AI agents is paramount for production deployment. Foundry prioritizes trust by embedding these controls directly into the platform, shifting the responsibility from individual developers to a comprehensive system-wide approach. The latest updates focus on post-build operations, providing mechanisms for monitoring agent performance, facilitating continuous improvement, and validating their value.
As agents scale from pilot programs to handling thousands of daily requests, Foundry offers tools to manage costs predictably without leaving the platform. This cost management is built on a foundation of choice, enabling organizations to select the most appropriate resources and configurations.
Key features for cost optimization include:
- Model Router: Intelligently matches each request to the most suitable model, balancing performance and cost.
- Prompt Caching: Reduces redundant computation by storing and reusing common prompts.
- PTU Spillover and Quota Optimization: Ensures service continuity during usage spikes by managing provisioned throughput units and optimizing resource allocation.
- Toolboxes in Foundry: Efficiently delivers only the necessary tools for each specific request, minimizing overhead.
- Agent Optimizer: Fine-tunes prompts, skills, tools, and model choices based on custom evaluators, enhancing both performance and cost-efficiency.
Beyond cost, Foundry provides a comprehensive view of ROI for agents, connecting business value, usage metrics, and operational costs. This unified dashboard allows teams to assess whether their production agents are delivering more value than they consume in resources, and to identify areas where cost may be outpacing benefits.
A Microsoft Mechanics episode on "Token Economics for Agents" offers a practical guide to understanding and optimizing these cost-saving mechanisms.
Real-World Impact: Organizations Building with Foundry
The adoption of Microsoft Foundry extends beyond experimentation, with organizations actively deploying agents to drive tangible business outcomes. From agile digital natives to established global enterprises, the trend is clear: teams that previously spent weeks integrating, securing, and deploying AI agents can now achieve these milestones in days. This accelerated deployment, coupled with infrastructure that meets stringent compliance standards and distribution through trusted user channels, is transforming how businesses leverage AI.
Notable examples of organizations building on Foundry include:
- Adobe: Enhancing customer experiences and streamlining creative workflows.
- Telefónica: Optimizing network operations and improving customer service.
- Tata Consultancy Services: Developing and deploying AI solutions for a diverse range of enterprise clients.
These companies, along with many others, are demonstrating the practical application and significant benefits of a unified, enterprise-grade AI agent platform.
Getting Started with Microsoft Foundry
All the features and capabilities discussed are now live and accessible within Microsoft Foundry. Comprehensive documentation and Microsoft Learn courses are available to guide users through the platform’s features. Developers can begin their journey with a quickstart guide, which provides an end-to-end walkthrough for setting up, testing, and deploying a production-ready hosted agent.
For those seeking structured learning, the "AI Agents for Beginners" curriculum offers a 12-lesson program. Deeper dives are available through guided labs such as "Develop AI Agents in Azure," "Hosted Agents Workshop (.NET)," the "Foundry Toolkit for VS Code and hosted agents workshop," and the "ZavaShop Supply Chain Workshop." To ensure the quality and reliability of deployed agents, the practical guide "Evaluating AI Agents: A Practical Guide with Microsoft Foundry" is also recommended reading.
A detailed explanation of Foundry Agent Service and the Microsoft Agent Framework can be found in a Microsoft Mechanics video featuring Jeff Hollan, offering insights into operationalizing AI agents from deployment to real-world impact.
