Microsoft AI CEO claims its model beats OpenAI’s GPT-5.6 using less compute: Why he said so

Microsoft AI CEO Mustafa Suleyman has claimed that the company's in-house artificial intelligence model outperforms the pre-trained base model powering OpenAI's GPT-5.6 Sol while using just a quarter of the computing resources.

Suleyman made the remarks during a town hall meeting with Microsoft employees on Thursday (October 15), according to a report by Business Insider. He said, according to the news outlet, Microsoft's model had performed better than OpenAI's in internal tests designed to evaluate how effectively AI systems understand and predict patterns in information.

"We now have a stronger pre-trained base model than the model that currently powers GPT 5.6 Sol," Suleyman said, according to Business Insider. "And in fact, we used one-quarter of the compute that OpenAI used to train their model."

The claims come even as Microsoft has substantial investment in OpenAI and the companies' longstanding commercial partnership.

Why is Microsoft comparing its AI model with OpenAI's?

Microsoft's push to develop proprietary AI models reflects its efforts to reduce dependence on OpenAI's technology and gain greater control over its AI products, development costs and computing infrastructure.

According to Business Insider, Suleyman said Microsoft had privately tested its model against OpenAI's offerings, including a model internally nicknamed "Spud", which powers GPT-5.6.

He added that Microsoft's comparisons were based on technical information the company had about OpenAI's models.

"Privately and confidentially, we've also compared it to the OpenAI models," Suleyman told employees, according to the report.

The comparison is significant because Microsoft is both a major investor in OpenAI and a commercial partner that uses its models across its products and cloud services.

By highlighting its model's performance against OpenAI's, Suleyman may be signalling Microsoft's ambition to compete more directly in the development of foundational AI technology.