ModelRefs / Chain-of-Thought — Prompt Pattern
Chain-of-Thought — Prompt Pattern
Force step-by-step reasoning before the final answer to lift accuracy on multi-step tasks.
Overview
Ask the model to think through the problem in explicit steps before giving the answer. Improves arithmetic, logic, and multi-hop reasoning at the cost of latency and tokens.
When to use it: You need accurate answers on multi-step problems and can afford extra tokens for reasoning.
Pattern details
| Pattern class | reasoning |
|---|---|
| Difficulty | beginner |
| Also known as | cot, step-by-step, let's think step by step |
| Last reviewed | 2026-06-07 |
Known failure modes
- Hallucinated steps — Model writes plausible but wrong intermediate steps. Mitigation: Pair with self-consistency or verifier prompt.
- Answer leaks into reasoning — Final answer appears in the middle of reasoning, breaking parsers. Mitigation: Use a strict 'Final answer:' fence and parse only after it.
When not to use it
- Asking for chain-of-thought on tasks that don't require reasoning (wastes tokens).
- Parsing the whole response as the answer.
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Chain-of-Thought — Prompt Pattern.
Frequently asked questions
When should I use the Chain-of-Thought prompt pattern?
You need accurate answers on multi-step problems and can afford extra tokens for reasoning.
What are common failure modes of Chain-of-Thought?
Hallucinated steps • Answer leaks into reasoning
Does Chain-of-Thought work across models?
Yes when the anatomy is preserved — see model compatibility matrix on the pattern page.