Microsoft's MAI-Base-2: 2.8T Parameters, a Quarter of the Compute
9 Articles
9 Articles
Microsoft's MAI-Base-2: 2.8T Parameters, a Quarter of the Compute
Microsoft has finished pre-training its largest base model to date. MAI-Base-2 is a mixture-of-experts system with 2.8 trillion total parameters, of which 80 billion are active for any given token. The company says it beats every open-weight model it tested, and that its training recipe is 4 to 12 times more compute-efficient than the one […]
Microsoft reveals cutting-edge MAI-Base-2 model with 2T parameters
Microsoft Corp. has developed a new large language model that it says is significantly more efficient than leading open-source alternatives. The company’s artificial intelligence research group detailed the algorithm on Thursday. MAI-Base-2 is not a full-fledged LLM but rather a base model, a kind of no-frills prototype. Developers turn base models into LLMs by putting […]
Microsoft AI Unveils MAI-Base-2, A 2.8-Trillion-Parameter Foundation Model with Superior FLOPs Efficiency
Microsoft AI has completed the pre-training of MAI-Base-2, a state-of-the-art Mixture-of-Experts (MoE) foundation model boasting 2.8 trillion total parameters and 80 billion active parameters. Trained on 16,000 NVIDIA GB200 GPUs over 38 days, the model demonstrates exceptional data efficiency, proving 4 to 12 times more FLOPs-efficient than leading open base models like DeepSeek-V4-Pro-Base across held-out […]
Microsoft's releases MAI-Base-2 aimed at open models with a twist
Microsoft's releases MAI-Base-2 aimed at open models with a twist Larry Dignan Fri, 16 Oct 2026 - 03:45 Larry Dignan Editor in Chief of Constellation Insights Constellation Research Larry Dignan is Editor in Chief of Constellation Insights at Constellation Research, where he leads editorial coverage focused on enterprise technology, digital transformation, and emerging trends shaping the future of business. He oversees research-driven news, ana…
Microsoft assures that MAI-Base-2 finished its pre-training after 38 days at 6,000 GPU and that it exceeds the open weight models evaluated. The company also highlights a computational efficiency of between four and twelve times that attributed to a DeepSeek comparison base.
Microsoft has announced that it has completed pre-training of its AI model "MAI-Base-2," which has a total of 2.8 trillion parameters. Read more...
Coverage Details
Bias Distribution
- 100% of the sources lean Left
Factuality
To view factuality data please Upgrade to Premium









