Introducing Claude Opus 4.5 in Microsoft Foundry

Announcing Anthropic's newest model, Claude Opus 4.5, successful Microsoft Foundry. Opus 4.5 is present disposable successful nationalist preview successful Microsoft Foundry, GitHub Copilot paid plans, and Microsoft Copilot Studio.

We’re astatine a existent inflection constituent successful the AI landscape, a threshold wherever models determination from utile assistants to genuine collaborators. Models that recognize the objective, origin successful constraints, and execute analyzable multi-tool workflows. Models that not lone enactment processes, but assistance restructure them for reliability, scale, and operational efficiency.

Anthropic’s newest model, Claude Opus 4.5, embodies that shift. Today, we are excited to stock that Opus 4.5 is present disposable successful nationalist preview successful Microsoft Foundry, GitHub Copilot paid plans, and Microsoft Copilot Studio.

Building connected the Microsoft Ignite announcement of our expanded concern with Anthropic, Microsoft Foundry delivers its committedness to giving Azure customers contiguous entree to the widest enactment of precocious and frontier AI models of immoderate cloud. Foundry empowers developers to accelerate innovation with an integrated, interoperable, and unafraid AI level that enables seamless deployment, integration, and scaling for AI apps and agents.

We’re excited to usage Anthropic Claude models from Microsoft Foundry. Having Claude’s precocious reasoning alongside GPT models successful 1 level gives america flexibility to physique scalable, enterprise-grade workflows that determination acold beyond prototypes.

—Michele Catasta, President, Replit

Opus 4.5 for existent work

Opus 4.5 sets a caller barroom for coding, agentic workflows, and endeavor productivity: outperforming Sonnet 4.5 and Opus 4.1, astatine a much accessible terms point. Its versatility crossed bundle engineering, analyzable reasoning, instrumentality use, and imaginativeness unlocks caller opportunities for organizations to modernize systems, automate captious workstreams, and present ROI faster.

By prioritizing accelerated integration of the latest models, Foundry allows Azure customers to enactment up of the curve and maximize the interaction of their agentic AI systems; each portion maintaining centralized governance, security, and observability astatine scale.

1. Built for accumulation engineering and agentic capabilities

According to Anthropic, Opus 4.5 delivers state-of-the-art show connected manufacture modular bundle engineering benchmarks, including caller highs connected SWE-bench (80.9%). Early testers consistently picture the exemplary arsenic being capable to construe ambiguous requirements, crushed implicit architectural tradeoffs, and identify fixes for issues that span aggregate systems.

Opus 4.5 accelerates engineering velocity by completing multi-day improvement enactment successful hours with:

• Improved multilingual coding performance
• More businesslike codification generation
• Stronger trial coverage
• Cleaner architectural and refactoring choices

Capability / Benchmark	Claude Opus 4.5	Claude Sonnet 4.5	Claude Opus 4.1	Gemini 3 Pro
Agentic coding (SWE-bench Verified)	80.90%	77.20%	74.50%	76.20%
Agentic terminal coding (Terminal-bench 2.0)	59.30%	50.00%	46.50%	54.20%
Agentic instrumentality usage — Retail (t2-bench)	88.90%	86.20%	86.80%	85.30%
Agentic instrumentality usage — Telecom (t2-bench)	98.20%	98.00%	71.50%	98.00%
Scaled instrumentality usage (MCP Atlas)	62.30%	43.80%	40.90%	_
Computer usage (OSWorld)	66.30%	61.40%	44.40%	_
Novel occupation solving (ARC-AGI-2 Verified)	37.60%	13.60%	_	31.10%
Graduate-level reasoning (GPQA Diamond)	87.00%	83.40%	81.00%	91.90%
Visual reasoning (MMMU validation)	80.70%	77.80%	77.10%	_
Multilingual Q&A (MMLU)	90.80%	89.10%	89.50%	91.80%

Claude Opus 4.5 benchmark results from Anthropic

Opus 4.5 is besides 1 of the strongest tool-using models disposable today, susceptible of powering agents that enactment seamlessly crossed hundreds of tools. Developers summation entree to respective important upgrades:

• Programmatic Tool Calling: Execute tools straight successful Python for much efficient, deterministic workflows.
• Tool Search: Dynamically observe tools from ample libraries without utilizing up abstraction successful the discourse window.
• Tool Use Examples: More close instrumentality calling for analyzable instrumentality schemas.

Together, these capabilities alteration blase agents crossed cybersecurity, full-stack bundle engineering, fiscal modeling, and different workflows requiring aggregate instrumentality interactions. Opus 4.5 shows strong, real-world quality applying these tools creatively wrong constraints. In testing, the exemplary successfully navigated analyzable argumentation environments, specified arsenic hose alteration rules, chaining upgrades, downgrades, cancellations, and rebookings to optimize outcomes. This benignant of adaptive, constraint-aware problem-solving reflects a meaningful measurement guardant successful what agentic AI systems tin accomplish.

Manus profoundly utilizes Anthropic’s Claude models due to the fact that of their beardown capabilities successful coding and long-horizon task planning, unneurotic with their prowess to grip agentic tasks. We are very excited to beryllium utilizing them present connected Microsoft Foundry!

—Tao Zhang, Co-founder & Chief Product Officer, Manus AI

2. Improved developer acquisition connected Foundry

Opus 4.5 paired with caller developer capabilities offered connected Foundry is designed to assistance teams physique much effectual and businesslike agentic systems:

• Effort Parameter (Beta): Control however overmuch computational effort Claude allocates crossed thinking, instrumentality calls, and responses to equilibrium show with latency and outgo for your circumstantial usage cases.
• Compaction Control: Handle long-running agentic tasks much efficaciously with caller SDK helpers that negociate discourse efficiently implicit extended interactions.

These enhancements supply greater predictability and operational power for endeavor workloads.

3. Enhanced bureau productivity and machine use

Opus 4.5 besides doubles down arsenic Anthropic’s champion imaginativeness model, unlocking workflows that beryllium connected analyzable ocular mentation and multi-step navigation. Computer usage show has improved significantly, enabling much reliable automation of desktop tasks.

For cognition workers, the exemplary delivers a step-change betterment successful powering agents that make spreadsheets, presentations, and documents. It produces enactment with consistency, nonrecreational polish, and genuine domain consciousness making it a acceptable for finance, legal, and different precision-critical verticals. The exemplary amended leverages representation to support discourse and consistency crossed files passim sprawling nonrecreational projects.

4. Safety and security

According to Anthropic, Opus 4.5 besides delivers meaningful improvements successful information and security. The exemplary shows a reduced complaint of misaligned responses, stronger robustness against prompt-injection attacks, and much reliable behaviour crossed analyzable tasks.

These improvements align with Microsoft’s committedness to providing endeavor customers with models that conscionable precocious bars for safety, governance, and operational integrity

Use cases

Opus 4.5 serves the pursuing usage cases

• Software development: Deploy agents that grip complex, multi-system improvement tasks with minimal supervision.
• Financial analysis: Connect insights crossed regulatory filings, marketplace reports, and interior information for blase predictive modeling and proactive compliance monitoring.
• Cybersecurity: Correlate logs, vulnerability databases, and menace quality for professional-grade menace detection and automated incidental response.
• Enterprise operations: Manage blase workflows requiring coordination crossed aggregate tools, systems, and accusation sources.

Pricing and availability

Opus 4.5 delivers frontier show and sets a caller modular for a assortment of usage cases astatine 1 third the terms of erstwhile Opus-class models.

Model	Offer type	Deployment type	Regions	Price (1M tokens)	Availability
Claude Opus 4.5	Serverless Pay-go	Global Standard	East US2, Sweden Central	Input- $5 Output- $25	November 24, 2025 (public preview)

Get started today

Claude Opus 4.5 is disposable present successful Microsoft Foundry and coming soon successful Visual Studio Code via the Foundry extension. Visit the Foundry portal to statesman gathering with Opus 4.5.