AI Glossary Term
What does “Chain-of-Thought” mean?
A reasoning style where an AI model decomposes a problem into intermediate steps.
How Chain-of-Thought works
Example
Asked a word problem, a model first writes out intermediate steps, such as subtracting then adding, before stating the final number. Research has shown that prompting for step-by-step reasoning can improve accuracy on math and logic problems compared with answering directly.
Why it matters
Showing steps can improve results on multi-step problems and lets people spot where an answer went wrong. The written steps, however, may not always faithfully reflect how the model reached its answer.
Common questions
Do I have to ask for chain-of-thought?
Often a simple instruction like 'think step by step' helps. Some newer models are trained to reason in steps automatically.
Does chain-of-thought guarantee a correct answer?
No. A model can write convincing steps that contain errors, so results still need checking.
Does it cost more?
Yes, usually. Writing out steps uses more tokens, which increases cost and response time.