Microsoft says MAI-Base-2 pre-training is complete

On Oct. 15, 2026 Microsoft AI said it finished pre-training MAI-Base-2, a mixture-of-experts model with 80 billion active parameters and 2.8 trillion total parameters. Microsoft said the run used 16,000 GB200 GPUs for 38 days in August and September 2026, and that on its own held-out tests the model beat every open-weight model it compared, with a claimed 4.2 to 12.3 times FLOPs-efficiency advantage over DeepSeek-V4-Pro-Base. The model is in post-training, with no release date or public weights.