Anthropic's most capable Opus model is now available where enterprise teams already build. Claude Opus 4.7 is available today on Microsoft Foundry, bringing frontier AI to the infrastructure, governance, and compliance workflows enterprises already rely on.
Anthropic continues to push the frontier of AI for real-world, production work with its new model Claude Opus 4.7. Opus 4.7 is Anthropic’s most advanced generally available model. Teams building on Azure can access Claude Opus 4.7 through Microsoft Foundry , today. This model is designed for the workflows enterprises run and delivers meaningful gains across agentic coding, long-running autonomous tasks, and professional work.
For organizations already building with Claude Opus 4.6, Opus 4.7 represents a powerful upgrade bringing stronger instruction following, better vision and office outputs, and improved memory capabilities to enable enhanced performance across complex, multi-step workflows..
What Microsoft Foundry brings to the table
Microsoft Foundry is Microsoft's unified platform for building, deploying, and governing AI applications at scale. It gives enterprise teams a single control plane for model selection, data connections, observability and access controls all backed by Azure's global infrastructure and security layer.
Bringing Claude Opus 4.7 into that environment means enterprises get frontier model capability without compromising governance properties their security and legal teams require.
Developers call Claude Opus 4.7 through Foundry's standard model APIs, which means existing tool chains, SDKs, and prompt harnesses require minimal changes. For teams already running Opus 4.6 in Microsoft Foundry, this is a direct upgrade path.
What Opus 4.7 adds for enterprise workloads
These improvements land directly in production workflows. Enterprises running Opus 4.6 today should see measurable gains in long-running tasks, coding, visual capabilities, instruction following, and reasoning. .
The model reasons through underspecified requests, making sensible assumptions and stating them clearly rather than stalling on clarifying questions a practical improvement for any workflow that touches real-world data. Check out Anthropic blog for benchmark table
Built-in enterprise governance
Running Opus 4.7 through Microsoft Foundry means every inference request inherits Foundry's enterprise controls: Azure Active Directory for identity and access, private networking via VNet and private endpoints, logging and audit trails through Azure Monitor.
Teams don't need to build a separate governance layer around the model. Foundry handles it, and security teams already know how to audit it.
Key use cases by vertical
Getting started
Claude Opus 4.7 is available today in Microsoft Foundry's model catalog. Teams already using Anthropic models on Microsoft Foundry can upgrade through the catalog with no infrastructure changes. New teams can deploy through the standard Foundry provisioning flow.
Claude Code is available via the Anthropic API routed through Microsoft Foundry; teams can point their existing Claude Code installations at their Foundry endpoint with a single configuration change.
GitHub Copilot CLI with Claude is also available to GitHub Copilot Enterprise subscribers today.
Anthropic has provided a migration guide covering the key differences between Opus 4.6 and Opus 4.7 to help teams get the most out of the new model with minimal harness changes.