menu
close

Microsoft Challenges OpenAI with New AI Reasoning Technology

Microsoft is developing proprietary AI reasoning models, codenamed MAI, to compete directly with OpenAI's advanced offerings. These models focus on enhanced logical reasoning and decision-making capabilities, with Microsoft planning to integrate them into Copilot and offer them as APIs to developers later in 2025. This strategic move signals Microsoft's intent to reduce its dependence on OpenAI despite their ongoing partnership.
Microsoft Challenges OpenAI with New AI Reasoning Technology

Microsoft is making significant strides in developing its own advanced AI reasoning models to compete with its longtime partner OpenAI, marking a major shift in the AI landscape.

The Redmond-based tech giant has completed training a family of models internally referred to as MAI, which reportedly perform nearly as well as leading models from OpenAI and Anthropic on commonly accepted benchmarks. According to sources familiar with the matter, Microsoft's AI division, led by Mustafa Suleyman, is focusing specifically on reasoning capabilities that use chain-of-thought techniques—a process that generates answers with intermediate reasoning steps when solving complex problems.

This development comes amid growing tensions between the two companies. Despite Microsoft's substantial $14 billion investment in OpenAI, reports indicate that OpenAI has been reluctant to share technical documentation about how its o1 reasoning model works, creating friction in the partnership. Microsoft has already begun experimenting with swapping MAI models for OpenAI's technology in its Copilot products and is considering releasing these models as application programming interfaces (APIs) later this year.

Beyond developing its own models, Microsoft is also testing AI models from competitors including xAI, Meta, Anthropic, and DeepSeek as potential alternatives to OpenAI's technology. This multi-model approach reflects Microsoft's broader strategy to diversify its AI capabilities and reduce its reliance on any single provider.

The timing is strategic, as Microsoft plans to invest $80 billion in AI-focused data centers in fiscal year 2025. Meanwhile, the company continues to advance its smaller Phi models, with the recent release of Phi-4 reasoning models demonstrating strong capabilities despite their smaller size.

"We are using a mix of models, which includes continuing our deep partnership with OpenAI, along with models from Microsoft AI and open source models," a Microsoft spokesperson stated, emphasizing that the company remains committed to providing choice in AI technologies while maintaining its partnership with OpenAI.

Source:

Latest News