ModelRefs / Extended Thinking — AI Glossary
Extended Thinking — AI Glossary
A model mode generating a long internal reasoning trace (scratchpad) before producing a final response, improving complex problem-solving.
Overview
Extended thinking (Claude 3.7 Sonnet 'thinking' mode, OpenAI o1/o3) allows the model to spend additional compute on internal chain-of-thought reasoning tokens before emitting the visible answer. The thinking process is often hidden from users. Dramatically improves performance on math, coding, and multi-step reasoning tasks at the cost of higher latency and token usage.
Reference details
| Topic | inference |
|---|---|
| Also known as | thinking mode, extended reasoning, scratchpad |
| Last reviewed | 2026-06-24 |
Related terms
Continue your research
Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Extended Thinking — AI Glossary.
Frequently asked questions
What is Extended Thinking?
A model mode generating a long internal reasoning trace (scratchpad) before producing a final response, improving complex problem-solving.
Is Extended Thinking the same as thinking mode?
Yes — thinking mode, extended reasoning, scratchpad are common aliases for Extended Thinking.
What concepts are related to Extended Thinking?
Closely related concepts include chain of thought, reasoning, tree of thought.