Which Claude models are generally available in Microsoft Foundry?

Claude Opus 4.8 and Claude Haiku 4.5 are the two models that hit GA in Microsoft Foundry on June 29, 2026. Both are served through the Anthropic Messages API and support prompt caching and extended thinking, so agentic and coding workloads carry over without rewriting the client.

How is Claude billed on Microsoft Foundry versus the Anthropic API?

Claude in Microsoft Foundry uses a single Azure invoice and, for customers with a Microsoft Enterprise Agreement, draws down their existing Azure commitment. Direct API calls to Anthropic still bill through Anthropic. Governance, identity and networking come from the customer's own Azure tenant on the Foundry path.

What are the two deployment options for Claude in Microsoft Foundry?

Claude in Microsoft Foundry offers hosted on Azure — inference in the Azure environment with a US data zone, Azure auth and Azure billing — or hosted on Anthropic, which exposes the full Messages API surface and any Claude model not yet on Azure. Teams pick per workload.

Which GPUs run Claude on Microsoft Foundry?

Claude in Microsoft Foundry runs on NVIDIA GB300 Blackwell Ultra GPUs in Azure, per NVIDIA and Anthropic. GB300 is NVIDIA's Grace Blackwell Ultra platform paired with Quantum-X800 InfiniBand, aimed at high-throughput inference for production AI agents.

Anthropic · 2026-06-29 · major

Claude in Microsoft Foundry — Opus 4.8 and Haiku 4.5 go GA on Azure

Claude Opus 4.8 and Claude Haiku 4.5 are now generally available in Microsoft Foundry on Azure, billed through the customer's Microsoft Enterprise Agreement and running on NVIDIA GB300 Blackwell Ultra GPUs.

Claude on Microsoft Foundry announcement banner combining Anthropic and Microsoft branding

Claude Opus 4.8 and Haiku 4.5 are now production-ready inside Azure, billed on the customer's Microsoft agreement.

Quick facts

Available models	Claude Opus 4.8, Claude Haiku 4.5
Interface	Anthropic Messages API inside Microsoft Foundry
Deployment options	Hosted on Azure or hosted on Anthropic
Billing	Draws down Microsoft Enterprise Agreement Azure commitment
Hardware	NVIDIA GB300 Blackwell Ultra GPUs
Data residency	US data zone available for teams with residency needs

What is it?

Microsoft Foundry — Azure's model catalog — moves Claude Opus 4.8 and Claude Haiku 4.5 from preview to general availability. Enterprises can call the Anthropic Messages API from inside their own Azure tenant, keeping Azure identity, networking and billing.

How does it work?

Two paths ship together. Hosted on Azure runs inference in the customer's Azure environment with a US data zone for data-residency workloads, backed by NVIDIA GB300 Blackwell Ultra GPUs. Hosted on Anthropic exposes the full Messages API and any Claude model not yet on Azure. Anthropic remains the data processor in both cases.

Why does it matter?

Microsoft Foundry GA gives Azure customers a first-party procurement, governance and networking path to Anthropic's flagship Claude Opus 4.8 without leaving their tenant. Combined with Enterprise-Agreement drawdown and GB300 acceleration, it removes the last commercial reasons enterprises were routing Claude through third parties.

Who is it for?

Enterprise Azure customers building agents and coding tools on Claude

Frequently asked questions

Which Claude models are generally available in Microsoft Foundry?: Claude Opus 4.8 and Claude Haiku 4.5 are the two models that hit GA in Microsoft Foundry on June 29, 2026. Both are served through the Anthropic Messages API and support prompt caching and extended thinking, so agentic and coding workloads carry over without rewriting the client.
How is Claude billed on Microsoft Foundry versus the Anthropic API?: Claude in Microsoft Foundry uses a single Azure invoice and, for customers with a Microsoft Enterprise Agreement, draws down their existing Azure commitment. Direct API calls to Anthropic still bill through Anthropic. Governance, identity and networking come from the customer's own Azure tenant on the Foundry path.
What are the two deployment options for Claude in Microsoft Foundry?: Claude in Microsoft Foundry offers hosted on Azure — inference in the Azure environment with a US data zone, Azure auth and Azure billing — or hosted on Anthropic, which exposes the full Messages API surface and any Claude model not yet on Azure. Teams pick per workload.
Which GPUs run Claude on Microsoft Foundry?: Claude in Microsoft Foundry runs on NVIDIA GB300 Blackwell Ultra GPUs in Azure, per NVIDIA and Anthropic. GB300 is NVIDIA's Grace Blackwell Ultra platform paired with Quantum-X800 InfiniBand, aimed at high-throughput inference for production AI agents.

Try it

Enable Anthropic in ai.azure.com and call the Messages API with your Azure credentials