Back to Table
ThThinking4

Thinking

AI that reasons step-by-step

modelsRow 4: Emergingintermediate2 hoursRequires: Lg

Overview

Thinking models explicitly reason through problems step-by-step, showing their work and improving accuracy on complex tasks.

What is it?

AI models that perform explicit reasoning before answering.

Why it matters

Chain-of-thought dramatically improves accuracy on math, logic, and multi-step problems. It makes AI reasoning more transparent.

How it works

Models are trained or prompted to break problems into steps. Extended thinking time allows more computation before the final answer.

Real-World Examples

o1/o3

OpenAI's reasoning models

Claude Extended Thinking

Anthropic's deep reasoning mode

DeepSeek-R1

Open-source reasoning model

Tools & Libraries

Chain-of-Thoughttechnique

Prompting technique for reasoning

Tree-of-Thoughttechnique

Exploring multiple reasoning paths

Self-Consistencytechnique

Sampling multiple reasoning chains