Chain-of-Thought (CoT) is a prompting and reasoning technique where an AI model explicitly works through intermediate reasoning steps before arriving at a final answer. Rather than jumping to a conclusion, the model breaks down complex problems into logical steps, dramatically improving accuracy on math, logic, coding, and multi-step reasoning tasks. CoT can be elicited through prompting ("think step by step") or trained into models directly. Extended thinking models like Claude with thinking, GPT o1/o3, and DeepSeek-R1 use internal CoT reasoning that may be hidden from the user. CoT is considered one of the most important discoveries in making LLMs more capable.
Frequently Asked Questions
What is chain-of-thought prompting?
Chain-of-thought prompting instructs an AI to reason step by step before answering, dramatically improving accuracy on complex tasks like math, logic, and multi-step problems.
What are thinking models?
Thinking models (like Claude with thinking and GPT o1) are trained to perform internal chain-of-thought reasoning automatically, using extended compute time to solve harder problems.