Microsoft Reveals 2 Trillion Parameter MAI-Base-2 Model

Oct 17, 2026
Microsoft unveiled MAI-Base-2, a mixture of experts base model with 2 trillion parameters that activates 80 billion parameters for each prompt.

Microsoft revealed MAI-Base-2, a base model with 2 trillion parameters, according to SiliconANGLE. The mixture of experts model activates 80 billion parameters for each prompt.

Microsoft says MAI-Base-2 outperformed DeepSeek-V4-Pro-Base on internal science, programming and math benchmarks. The company also says the model is up to 12 times more efficient on certain hardware measures.

Microsoft trained MAI-Base-2 for 38 days using 16,000 Nvidia GB200 chips and a dataset of 50 trillion tokens. About one fifth of the tokens trained its computer vision module, which processes material such as scientific charts.

The company will license the model weights to trusted partners through Azure. It also plans to use MAI-Base-2 as the foundation for additional language models.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Daily AI Brief.

Also, consider following us on social media:

Free newsletter

Daily AI Brief

Daily report covering major AI developments and industry news, with both top stories and complete market updates

Industry analysis

2025 Global Business Services Agenda: Gen AI Takes Center Stage

The Hackett Group

This industry analysis by The Hackett Group explores the transformative impact of generative artificial intelligence (Gen AI) on global business services (GBS) in 2025. The study highlights the shift from exploration to acceleration of Gen AI initiatives, with 89% of executives advancing these projects to improve customer satisfaction, innovate products, and reduce costs. The report also discusses the challenges and strategies for successful Gen AI adoption, emphasizing the need for a technology-enabled operating model and the importance of reskilling the workforce.

Read more
Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.