Reasoning model
추론 모델
An AI model built to work through a problem step by step before it answers, instead of answering right away.
Overview
An ordinary LLM starts writing the answer as soon as it gets a question. A reasoning model first goes through a "thinking" stage. It breaks the problem down, tries several approaches and checks itself, and then gives its answer.
It is strong at jobs that need careful working-out, such as math problems, complex coding and plans with many conditions tangled together. The trade-off is that answers are slower and cost more because it uses more tokens. So many services let you choose how deeply it should think.
An everyday example
It is the difference between answering a sum in your head right away and writing out the working on paper and checking it.
Why it matters
Using a fast model for easy questions and a reasoning model for hard problems has become the basic way to use AI today.
Also calledthinking model