Microsoft's AI chief claims company outperformed OpenAI with less computing power

Microsoft's AI chief, Mustafa Suleyman, told employees that the company's in-house model, MAI-Base-2, outperformed OpenAI's offerings, including the GPT-5.6 model, using just a quarter of the computing power. Suleyman stated during a town hall meeting that Microsoft's model had a stronger pre-trained base than OpenAI's, and that it performed better in pattern recognition tests. The model was announced earlier on Thursday. Suleyman also noted that Microsoft's model used 12 times less computation than DeepSeek's, according to a public blog post. Microsoft has invested billions in OpenAI and owns a stake likely worth over $100 billion. The relationship between the two companies has become strained as Microsoft increasingly focuses on its own AI models. Microsoft declined to comment, and OpenAI did not respond to a request for comment.