Microsoft 365 Copilot Expands with Anthropic’s Claude Models
Key Points
- Microsoft adds Anthropic’s Claude Sonnet 4 and Claude Opus 4.1 to 365 Copilot.
- Claude models complement existing ChatGPT integration, not replace it.
- Both models are available in Copilot’s Researcher agents and Copilot Studio.
- Access is limited to customers who opt in via the Frontier Program.
- Admins can enable the models through the Office 365 admin center.
- Users can compare Claude and ChatGPT outputs for specific tasks.
- The rollout aims to provide diverse AI perspectives within the same platform.
Microsoft announced that its AI assistant, 365 Copilot, will now include two Anthropic models—Claude Sonnet 4 and Claude Opus 4.1. The new models will sit alongside OpenAI’s ChatGPT, offering users alternative AI perspectives for specific tasks. They will be available in Copilot’s Researcher agents and Copilot Studio, with access initially limited to customers who opt in through the Frontier Program.
New Model Additions
Microsoft disclosed that 365 Copilot is adding two of Anthropic’s large‑language models: Claude Sonnet 4 and Claude Opus 4.1. These models complement, rather than replace, the existing OpenAI‑based ChatGPT integration. By incorporating Anthropic’s technology, Copilot can deliver varied responses, reflecting differences in training data and model architecture.
Placement in Copilot Features
The Claude models will be integrated into two core Copilot components. First, they will appear as selectable options within the Researcher agents, which are reasoning agents that draw on daily data such as email, calendar, and files to generate ideas, reports, and product concepts. Second, both Claude models will be usable in Copilot Studio, a tool that lets organizations build custom enterprise agents using a mix of OpenAI and Anthropic models.
Access and Availability
Access to the new models is currently limited. 365 Copilot customers with appropriate licenses can opt in through Microsoft’s Frontier Program to enable Claude Opus 4.1 in Researcher agents. Organizations wishing to build agents with Claude in Copilot Studio must also opt in. Administrators can activate the models via the Office 365 admin center once the opt‑in process is complete.
Implications for Users
By offering Anthropic’s models alongside ChatGPT, Microsoft provides Copilot users with the ability to compare outputs from different AI systems and select the one that best fits their workflow. This diversification may improve the quality of generated content, especially in scenarios where model‑specific strengths are advantageous. The limited rollout through the Frontier Program suggests Microsoft is gathering early feedback before broader deployment.