ITG NowHome › AI glossary › Mixture of experts
What is a mixture of experts model?
Short answer
Mixture of experts is an AI design where a model has many specialist parts and uses only a few for each task.
- Category Models
- Also called MoE
- Updated 20 Oct 2026
Mixture of experts, explained simply
Imagine a hospital with many specialists. A front desk sends you only to the ones you need.
A mixture of experts model works the same way. A small router picks which parts of the model to use for each piece of text.
This lets a model be big in total but cheaper to run, since most of it rests at any moment.
On ITG Now, our AI models page and news stories may note when a model uses a mixture of experts design.
Related terms
TransformerParametersLLMInferenceAI model
Where you'll see it on ITG Now
- GuideColibri: Run Large AI Models Locally Without a GPU
- AI newsMicrosoft builds a giant 2.8 trillion parameter AI
- AI newsNew US lab gives away a giant AI free
- AI newsReflection Beam: open coding model arrives on AWS
Questions people ask about Mixture of experts
Is mixture of experts a type of model?
It is a way of building a model. Many LLMs can use this design.
Why do people use mixture of experts?
It can give the skill of a big model at a lower running cost.