OpenAI’s GPT-5.5 is generally available on Microsoft Foundry, bringing OpenAI’s latest frontier model to Azure for enterprise teams to build agents for real-world production work.
OpenAI’s GPT-5.5 will be publicly available tomorrow at Microsoft Foundrybring OpenAI’s modern frontier models to Azure and help enterprise teams build agents for real-world production work.
GPT-5.5 continues the clear progression of the GPT-5 series. GPT-5 brings unified reasoning and speed to a single system. GPT-5.4 brings more powerful multi-step inference and early agent capabilities for enterprise use. GPT-5.5 advances this arc with deeper long-context inference, more reliable agent execution, improved computational precision, and increased token efficiency, designed for persistent, high-stakes professional workflows.
Powerful models alone are not enough to operate agent AI at scale. microsoft foundry provides a platform layer that turns frontier models into usable and manageable systems, allowing enterprises to apply security policies and controls at the platform level. Foundry is a unified, interoperable environment to confidently build, optimize, and deploy AI applications and agents. Customers benefit from a wide selection of models, an open and flexible agent framework, native integration with enterprise systems and productivity tools, and enterprise-grade security, compliance, and governance. As new models like GPT-5.5 become available, Foundry enables them to be smoothly evaluated, commercialized, and scaled.
New features in GPT-5.5
GPT-5.5 is built for professional scenarios where accuracy, reliability, and durability are important. The premium variant, GPT-5.5 Pro, extends inference depth and task complexity for the most demanding enterprise workloads.
- Improving agent coding and computer usage: Execute multi-step engineering tasks end-to-end. This means maintaining context across large systems, diagnosing the root cause of ambiguous faults at the architectural level, and reasoning about how a fix affects other parts of the codebase before making a change. Anticipate downstream test and review requirements without being prompted and navigate the software interface with increased accuracy and more reliable recovery when execution goes in an unexpected direction.
- Autonomous execution and depth of research: Go beyond code to handle the full range of professional work and create sophisticated deliverables such as documents, spreadsheets, and presentations. For research-intensive workflows, GPT-5.5 acts as an active collaborator throughout the arc from question to output. That means refining drafts across multiple passes, stress-testing analytical reasoning, suggesting approaches, and integrating documentation, data, and code to drive the work, not just answer it.
- Complex reasoning and long context analysis: Handle extensive documentation, codebases, and multi-session history without losing threads.
- Token efficiency built with scale in mind: GPT-5.5 delivers high-quality output with fewer tokens and fewer retries, reducing costs and latency for large-scale production deployments.
GPT-5.5 is particularly suited to fields where the cost of inaccuracy is high, such as software engineering, DevOps, law, health sciences, and professional services. Microsoft Foundry’s GPT-5.5 allows customers to combine OpenAI’s modern frontier model with enterprise-grade infrastructure to bring agent AI into production.
Microsoft Foundry: Operating system for large-scale GPT-5.5 agents
Access to frontier models is just the starting point. What we’ve seen from our customers is that the hard part isn’t building the agents, it’s running thousands of agents in a production environment with real separation, identity, and governance. That’s where Foundry Agent Service comes into play.
There’s a revolution in the market. Now, developers can reliably infer business problems using a coding agent (a human interacting with a model to think, explore, and ask questions) and whose output becomes a production agent, a declarative workflow appropriate for a specific task and connected to business systems.
These declarative agents can be defined in YAML or written in a harness such as Microsoft Agent Framework, GitHub Copilot SDK, or virtually any library. LangGraph, Claude Agent SDK, and OpenAI Agents SDK all work the same way when using Foundry Agent Service hosted agents. Engineers can run a single command to deploy agents in an isolated sandbox with a persistent file system, clear Microsoft Entra ID, and zero-scale pricing. Large-scale, enterprise-ready agent powered by GPT-5.5.
Learn more about Foundry Agent Service and Microsoft Agent Framework.
Pricing
| model | input ($/M tokens) | Cached input ($/M tokens) | output ($/M tokens) |
| GPT-5.5 | $5.00 | $0.50 | $30.00 |
| GPT-5.5 Pro | $30.00 | $3.00 | $180.00 |
Try using Microsoft Foundry
Stop experimenting and start building your next production AI workload with GPT-5.5 from Microsoft Foundry.
