---
title: "Microsoft finishes training MAI-Base-2, a 2.8-trillion-parameter model it says needs up to 12 times less compute than DeepSeek’s"
url: https://anawords.com/story/microsoft-mai-base-2-pretraining-2-8-trillion-parameters
published: 2026-10-16T15:37:48+05:30
updated: 2026-10-16T15:37:48+05:30
section: Models & Launches
author: Mayank Sahu
checked: sources checked by AI, approved by editor
version: 1
---

# Microsoft finishes training MAI-Base-2, a 2.8-trillion-parameter model it says needs up to 12 times less compute than DeepSeek’s

The in-house foundation model is still being post-trained. Microsoft plans to license its weights to trusted partners through Foundry rather than release them openly.

**In short:** Microsoft AI said on 15 October 2026 that it has finished pre-training MAI-Base-2, a mixture-of-experts model with 2.8 trillion total and 80 billion active parameters. Microsoft claims it is 4.2 to 12.3 times more compute-efficient than DeepSeek-V4-Pro-Base. It is still in post-training.

## What Microsoft announced

Microsoft AI, the division led by Mustafa Suleyman, said on 15 October 2026 that it has completed pre-training of MAI-Base-2, its next in-house foundation model.[^1][^2] It is a mixture-of-experts model with 2.8 trillion total parameters and 80 billion active.[^1][^2]

## How it was trained

According to Microsoft:[^1][^3]

- Training ran for 38 days in August and September 2026 on 16,000 Nvidia GB200 GPUs.
- It used a 50-trillion-token training mix, with images making up 18%, so vision is built in from the start.
- The model was trained on original data rather than by distilling other companies’ models.

## What Microsoft claims

Microsoft says MAI-Base-2 is 4.2 to 12.3 times more compute-efficient in pre-training than DeepSeek-V4-Pro-Base across four evaluations, and outperforms every open-weight model it tested.[^1][^2] It says the training recipe is 6.3 to 12.7 times more efficient than its predecessor, MAI-Base-1.[^1] These are Microsoft’s own comparisons, made against base and post-trained checkpoints that it notes may not be like for like.[^1]

## When and how it will be offered

MAI-Base-2 is still in post-training, and Microsoft gave no release date.[^1][^2] It plans to license the model weights to trusted partners through Foundry.[^1][^2] GIGAZINE notes the earlier MAI-Base-1 was closed-weight.[^3] Microsoft says work on MAI-Base-3 has begun.[^1]

## Why it matters

Microsoft is building its own frontier models alongside its partnership with OpenAI. It is also building its own chips; see our story on [Maia 300](/story/microsoft-maia-300-inference-chip).

## Fact ledger

- [Confirmed] MAI-Base-2 has 2.8 trillion total and 80 billion active parameters. (source 1,2)
- [Confirmed] It was trained for 38 days on 16,000 Nvidia GB200 GPUs. (source 1,2)
- [Company claim] It is 4.2 to 12.3 times more compute-efficient than DeepSeek-V4-Pro-Base. (source 1,2)
- [Confirmed] Microsoft plans to license the weights to trusted partners through Foundry. (source 1,2)
- [Not known] Release date and pricing.

## Sources

1. [Microsoft AI: Frontier pre-training at Microsoft AI](https://microsoft.ai/news/frontier-pre-training-at-microsoft-ai/) — official
2. [Investing.com (via Yahoo Finance): Microsoft AI unveils 2.8 trillion-parameter model](https://finance.yahoo.com/technology/ai/articles/microsoft-ai-unveils-2-8-165617565.html)
3. [GIGAZINE: Microsoft completes pre-training of MAI-Base-2](https://gigazine.net/gsc_news/en/20261016-microsoft-mai-base-2/)