ع
Start Topics Teams Reference What's new Saved
AI & models

Reasoning

Chain of thought · Extended thinking

Reasoning is when a model works through a problem step by step before giving its final answer, instead of blurting out the first thing that fits. Laying out the intermediate steps — like showing your work on a math problem — meaningfully improves accuracy on hard, multi-step questions where a snap answer would trip up.
Why it matters

It's why 'thinking' or 'extended thinking' modes get math, logic, and tricky planning right far more often — the trade-off is they're slower and use more tokens, so you save them for the hard stuff.

Part of How hard Claude thinks

← All terms