Microsoft Claims New MAI-Base-2 AI Model Beats OpenAI's GPT-5.6 on Less Compute

Microsoft AI CEO Mustafa Suleyman told employees during a Thursday town hall that the company's new MAI-Base-2 model is stronger than the model currently powering OpenAI's GPT-5.6 Sol. Suleyman stated the pre-trained base model was built using just 1/4 of the compute power that OpenAI required. Microsoft privately tested the new model against OpenAI's models, including a model nicknamed "Spud" that powers GPT-5.6, using technical information the company accesses confidentially. The model also used 12x less computation than DeepSeek-V4-Pro-Base, according to Suleyman and a public blog post detailing the pre-training efficiency. Microsoft has invested billions into OpenAI and owns a stake in the startup likely worth more than $100B, while simultaneously using OpenAI models for Azure cloud customers and internal Copilot products. "We now have a stronger pre-trained base model than the model that currently powers GPT 5.6 Sol. And in fact, we used one-quarter of the compute that OpenAI used to train their model." — Mustafa Suleyman