Microsoft AI says in-house MAI-Base-2 finished training and beats open-weight models

1. 2:17 PM
   🚨 AI News | TestingCatalog@testingcatalogX

   Microsoft completes MAI-Base-2 pre-training, outperforming open-weight models

   AIMicrosoft says MAI-Base-2 outperforms every open-weight model it tested on held-out evaluations. The company says the model is 4-12x more pre-training FLOPs-efficient than DeepSeek-V4-Pro-Base, the strongest open base model, and that it is using those savings to train a larger model. Mustafa Suleyman says MAI-Base-2 is a 3T-parameter-class base model built entirely in-house, with post-training results to follow.

   
2. 12:16 PM
   Mustafa Suleyman@mustafasuleymanX

   Microsoft's MAI-Base-2 completes full training run without issues

   AIMustafa Suleyman says MAI-Base-2 trained cleanly across its full run. He credits careful scaling ladders and steady refinement of the training recipe.

   
3. 12:16 PM
   Mustafa Suleyman@mustafasuleymanXPick

   Microsoft AI says its in-house 3T-parameter base models are state of the art

   AIMustafa Suleyman says MAI now has state-of-the-art pre-trained base models in the 3T-parameter weight class, built entirely in-house. He says post-training results are coming soon and points readers to a technical blog post for more details.

The trend compares only continuously observed participants, so its scope may be smaller than the current heat total. Hover or tap for hourly heat; use arrow keys with a keyboard.