Reasoning models aim to tackle tasks that require step-by-step logical thinking. They often integrate planning or structured representations to enable complex problem solving.

Key Ideas

• Break down a problem into intermediate steps
• Use external tools or memory to keep track of progress
• Chain-of-thought prompts guide large language models to reason more effectively

Simple Example

A reasoning approach might ask a language model to first outline a solution, then verify each step before returning a final answer. This approach can produce more reliable results compared to a single-shot response.