BitTide
中文 Sign in

mai-thinking-1

Tag

Cards List
#mai-thinking-1

@Datou: Microsoft values its reputation, deliberately avoiding synthetic data. They trained a base model using only human data, then split it into three expert models for different domains. They then distilled these three capabilities back into the base model (weight ratio allocation requires experience), followed by a round of reinforcement learning to enable the distilled model to flexibly apply the right capability based on the problem.

X AI KOLs Timeline ↗ · 2026-06-02 Cached

Microsoft releases technical details of MAI-Thinking-1 training: uses purely human data to train a base model, then trains three domain expert models, merges capabilities back into the base model via distillation, and then applies reinforcement learning to enable the model to flexibly utilize different capabilities.

0 favorites 0 likes
#mai-thinking-1

@yvbbrjdr: I recommend everyone to read the MAI-Thinking-1 technical paper. It contains detailed (almost all) information on how to train a SOTA LLM. https://microsoft.ai/wp-content/uploads/2026/06/ma…

X AI KOLs Timeline ↗ · 2026-06-02 Cached

Recommended reading: the MAI-Thinking-1 technical paper, which details almost all the steps to train a SOTA large language model.

0 favorites 0 likes
#mai-thinking-1

Microsoft’s first advanced reasoning AI is here

The Verge ↗ · 2026-06-02 Cached

Microsoft announced MAI-Thinking-1, a flagship reasoning AI model, alongside six other new models at Build 2026, marking a major step in in-house model development.

0 favorites 0 likes
← Back to home

Submit Feedback