Reasoning
reasoning models, Reasoning models
A model's ability to work through a problem in intermediate steps instead of answering straight away. Reasoning models generate their own “thinking tokens” (chain of thought) before the actual answer, check approaches and correct themselves.
💭 Imagine …
Working it out on paper instead of in your head: slower, but much better on hard problems.
🔎 Example
Multi-step calculations, debugging code or planning a data migration. Many assistants show that they are “thinking” while doing it.
🎯 Why it matters
Thinking tokens cost time and money like any other output. Worth it for tricky tasks, not for simple ones. How much thinking happens is controlled by model effort.
🎓 Prefer step by step?
The AI Literacy course: 6 short modules with quizzes, a solid foundation in about 60 minutes.