OpenAI is no longer Microsoft’s only child.
On Wednesday, Microsoft announced that it’s adding Anthropic’s models to its Copilot Studio. Users can now choose between Anthropic’s Claude Sonnet 4 or Opus 4.1 and OpenAI’s GPT-4o.
Anthropic’s models, launched Wednesday in early release cycle environments, will fully roll out in the next two weeks.
- To start, users will be able to leverage Anthropic’s Claude Opus 4.1 for research tasks.
- Additionally, Opus 4.1 and Claude Sonnet 4 will be available to create and customize “enterprise-grade” agents.
- “And stay tuned: Anthropic models will bring even more powerful experiences to Microsoft 365 Copilot,” Charles Lamanna, president of business and industry for Copilot, wrote in a blog post.
Though Microsoft and OpenAI still walk arm-in-arm, bringing rival Anthropic into the mix could signal that the company is seeking to broaden its horizons.
Microsoft and OpenAI’s partnership first began in 2019 when the company invested $1 billion in the startup, followed by an additional $10 billion investment in 2023. The move united two of AI’s power players when the race was first heating up, and allowed Microsoft to carve out a significant niche in AI for the workplace, powered by OpenAI’s models.
That relationship has since grown tense as OpenAI has skyrocketed in popularity, and reached a boiling point when OpenAI tried (and failed) to acquire AI coding platform Windsurf in June. The waters have settled in recent weeks, with the two reaching a tentative agreement to revise the terms of their partnership which would allow the startup to restructure itself.
Microsoft, too, has been working on beefing up its own in-house models. Earlier this month consumer AI chief Mustafa Suleyman said the company was making “significant investments” in its own infrastructure to train AI.