OpenAI’s GPT-5.5 volition beryllium mostly disposable successful Microsoft Foundry, bringing OpenAI’s latest frontier exemplary to Azure and the endeavor teams gathering agents for existent accumulation work.
OpenAI’s GPT-5.5 volition beryllium mostly disposable time successful Microsoft Foundry, bringing OpenAI’s latest frontier exemplary to Azure and the endeavor teams gathering agents for existent accumulation work.
GPT-5.5 continues a wide progression successful the GPT-5 series. GPT-5 brought unified reasoning and velocity into a azygous system. GPT-5.4 brought stronger multi-step reasoning and aboriginal agentic capabilities for endeavor use. GPT-5.5 advances this arc with deeper long-context reasoning, much reliable agentic execution, improved computer-use accuracy, and greater token efficiency—designed for sustained, high-stakes nonrecreational workflows.
Powerful models unsocial aren’t capable to operationalize agentic AI astatine scale. Microsoft Foundry provides the level furniture that turns frontier models into usable, governable systems that alteration enterprises to use information argumentation and absorption astatine the level level. Foundry is simply a unified, interoperable situation to build, optimize, and deploy AI applications and agents with confidence. Customers payment from wide exemplary choice, unfastened and flexible cause frameworks, autochthonal integration with endeavor systems and productivity tools, and enterprise-grade security, compliance, and governance. When caller models similar GPT-5.5 go available, Foundry makes it casual to evaluate, productionize, and standard them without friction.
What’s caller successful GPT-5.5
GPT-5.5 is built for nonrecreational scenarios wherever precision, reliability, and persistence matter. GPT-5.5 Pro, a premium variant, extends reasoning extent and task complexity for the astir demanding endeavor workloads.
- Improved agentic coding and computer-use: Executes multi-step engineering tasks end-to-end—holding discourse crossed ample systems, diagnosing the basal origin of ambiguous failures astatine the architectural level, and reasoning done what other successful the codebase a hole volition impact earlier making a move. It anticipates downstream investigating and reappraisal requirements without needing to beryllium told, and navigates bundle interfaces with improved precision and much reliable betterment erstwhile execution takes an unexpected turn.
- Autonomous execution and probe depth: Goes beyond codification to grip the afloat span of nonrecreational work—producing polished deliverables similar documents, spreadsheets, and presentations. For research-intensive workflows, GPT-5.5 operates arsenic an progressive collaborator crossed the full arc from question to output: refining drafts crossed aggregate passes, stress-testing analytical reasoning, proposing approaches, and synthesizing crossed documents, data, and codification to thrust enactment guardant alternatively than conscionable answering it.
- Complex reasoning and long-context analysis: Handles extended documents, codebases, and multi-session histories without losing the thread.
- Token ratio built for scale: GPT-5.5 reaches higher-quality outputs with less tokens and less retries—lowering outgo and latency for accumulation deployments astatine scale.
GPT-5.5 is peculiarly good suited for domains wherever the outgo of imprecision is high—such arsenic bundle engineering, DevOps, legal, wellness sciences, and nonrecreational services. With GPT-5.5 successful Microsoft Foundry, customers tin brace OpenAI’s latest frontier exemplary with enterprise-grade infrastructure to enactment agentic AI into production.
Microsoft Foundry: The operating strategy for GPT-5.5 agents astatine scale
Access to a frontier exemplary is conscionable the starting point. What we spot from customers is that the hard portion isn’t gathering an agent: it’s moving thousands of them successful production, with existent isolation, identity, and governance. That’s wherever Foundry Agent Service comes in.
A gyration is unfolding successful the market. Now, a developer tin reliably crushed done a concern occupation with a coding agent—human’s interact with a exemplary doing dense thinking, research, asking questions — and the output is simply a accumulation agent: a declarative workflow suitable for a circumstantial task, and connected to your concern systems.
These declarative agents tin beryllium defined successful YAML oregon written successful a harness similar Microsoft Agent Framework, GitHub Copilot SDK, oregon virtually immoderate library. With hosted agents successful Foundry Agent Service, LangGraph, Claude Agent SDK, and OpenAI Agents SDKs each enactment the aforesaid way. Engineers tin tally a azygous bid to onshore agents successful an isolated sandbox with a persistent filesystem, a chiseled Microsoft Entra identity, and scale-to-zero pricing. Enterprise acceptable agents, astatine scale, powered by GPT-5.5.

Learn much astir 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 |
Get started with Microsoft Foundry
Stop experimenting and commencement gathering your adjacent accumulation AI workloads with GPT-5.5 successful Microsoft Foundry.