TMTPOST -- Microsoft Corporation is take further steps to pivot from OpenAI’s artificial intelligence (AI) technology.

Credit：Microsoft

Microsoft on Wednesday said it began to offer new AI model options from startup Anthropic for the Microsoft 365 Copilot assistant. That is the first time for Microsoft to allow ists business customers to access models from an OpenAI’s rival in its AI productivity assistant.

Starting Wednesday, Anthropic's Claude Sonnet 4 and Claude Opus 4.1 models showing up in Microsoft 365 Copilot, specifically, in Researcher, a reasoning agent, and Copilot Studio, an end-to-end conversational AI platform. 

Researcher help tackle complex, multistep research—reasoning over not only the web and trusted third-party data, but all of the work content across users' emails, chats, meetings, files, and more to deliver highly-skilled expertise on demand.With multiagent systems and prompt tools in Copilot Studio, customers can mix which models are used for specialized tasks from Anthropic, OpenAI, or any of the models in the Azure Model Catalog.

Microsoft said Copilot will continue to be powered by OpenAI’s latest models, and now customers have the flexibility to use Anthropic models too. Once opt-in, cutomers will be able to switch between OpenAI and Anthropic models in Researcher. In Copilot Studio, a drop-down menu lets  customers select from a range of model options—no more jumping between platforms to find the right tool.

“The addition of Claude Sonnet 4 and Claude Opus 4.1 advances our commitment to bring the best AI innovation from across the industry to Microsoft 365 Copilot, tuned for work and tailored to your business needs,” said Charles Lamanna, president of Microsoft’s business and industry Copilot team, in a blog post.
Lamanna suggested promising adoption of Anthropic models. “This is just the beginning—we’re committed to delivering model innovation at speed to empower every person with Microsoft 365 Copilot and transform every business process with agents,” he said.  “And stay tuned: Anthropic models will bring even more powerful experiences to Microsoft 365 Copilot.”

As of last October, Microsoft has invested $13.75 billion in OpenAI since 2019, including its share of the startup’s $6.6 billion fundraising in October. Such massive investment entitles Microsoft to a percentage of OpenAI’s profits. 

The partnership between OpenAI and Microsoft has grown more complex as both companies have expanded their AI ambitions. Microsoft has been developing its own AI models, and in a filling in August 2024, the tech giant for the first time called OpenAI a “competitor”, adding the startup to a long list of AI competitors, alongside Anthropic, Amazon and Meta. OpenAI was also listed alongside Google as a competitor to Microsoft in search, thanks to OpenAI’s SearchGPT feature announced in late July 2024.

OpenAI in the summer of 2024 started tapping Oracle for compute. In January, Stargate, OpenAI’s massive new AI infrastrucure deal with Oracle and SoftBank was announced. The same month saw Microsoft would not no longer be the exclusive provider of data center infrastructure for OpenAI to train and run its AI models.

Under a new agreement, Microsoft gets first choice over whether to host OpenAI’s AI workloads in the cloud — but if Microsoft can’t meet its needs, OpenAI can go to a rival cloud provider. OpenAI in June was reported to sign a cloud deal with Google despite the two companies’ AI rival.

OpenAI has signed a deal to purchase $300 billion worth of computing power from Oracle over roughly five years, marking one of the biggest cloud contracts ever signed, the Wall Street Journal on September 10 reported, citing people familiar with the matter. The contract, starting in 2027, requires 4.5 gigawatts (GW) of power capacity, , roughly comparable to the electricity amount consumed by about four million homes, per the report.