Microsoft is creating its own AI reasoning models to challenge OpenAI in the competitive artificial intelligence space.
Microsoft is reportedly working on developing its own artificial intelligence (AI) reasoning models to rival OpenAI and plans to potentially offer them to developers, according to a recent report from The Information. This move is part of Microsoft's ongoing efforts to reduce its reliance on OpenAI, despite being one of the startup's largest backers.
The company, based in Redmond, Washington, has begun experimenting with AI models from third-party firms, including xAI, Meta, and DeepSeek, as potential replacements for OpenAI's technology in its flagship product, Microsoft 365 Copilot. This shift comes as Microsoft seeks greater flexibility and cost control over its AI infrastructure.
Microsoft's partnership with OpenAI helped propel the company to the forefront of the competitive AI race among major tech players, but now the tech giant is working to diversify its AI portfolio. In December, Reuters exclusively reported that Microsoft had been exploring the integration of both internal and third-party AI models to power its Microsoft 365 Copilot, aiming to move away from sole reliance on OpenAI's GPT technology.
When Microsoft introduced Microsoft 365 Copilot in 2023, one of its major selling points was the use of OpenAI's GPT-4 model. However, the new report indicates that Microsoft's AI division, led by Mustafa Suleyman, has successfully trained a new family of models known internally as "MAI." These models are said to perform nearly as well as the top models from OpenAI and Anthropic on established benchmarks.
The MAI models are part of a broader effort by Microsoft to develop reasoning models that use "chain-of-thought" techniques. These techniques enable the AI to generate answers through a process of intermediate reasoning, which could allow the models to solve more complex problems. The reasoning capabilities are expected to compete directly with OpenAI's models.
Suleyman's team is already testing the integration of these MAI models, which are significantly larger than a previous family of Microsoft models called Phi, into Microsoft 365 Copilot as a potential replacement for OpenAI's technology. Furthermore, Microsoft is considering releasing the MAI models as an application programming interface (API) later this year, enabling external developers to integrate them into their own applications.
This development highlights Microsoft's continued ambition to establish itself as a major player in the AI space, with a focus on innovation, independence, and cost management. The company’s move to create and potentially monetize its own AI models signals a new chapter in the rapidly evolving AI landscape.