ITG NowHome › AI glossary › Distillation
What is distillation in AI?
Distillation is a way to teach a small AI model to copy the answers of a bigger one.
- Category Models
- Also called model distillation, distilled model
- Updated 20 Oct 2026
Distillation, explained simply
A big model is like a skilled teacher. A small model is the student that learns from the teacher's answers.
The student ends up faster and cheaper to run, and often keeps much of the teacher's skill.
That is how some small models can run on a phone and still give useful replies.
On ITG Now, our AI news stories and model comparisons sometimes mention distilled models that are small but capable.
Related terms
Small language modelTrainingAI modelLLMInferenceQuantization
Where you'll see it on ITG Now
- AI newsHackers breach OpenCode, run up AI bills
- AI newsUS bars DeepSeek from civilian government agency systems
Questions people ask about Distillation
Distillation vs quantization: what is the difference?
Distillation trains a smaller model to copy a bigger one. Quantization shrinks a model by using simpler numbers.
Is distillation allowed?
It depends on the license and terms of the model you copy from. Check the rules first, and this isn't legal advice.