Reasoning models

Reasoning models are AI models that work through a problem step by step before answering, like OpenAI's o-series, DeepSeek-R1 and Gemini's thinking models.

AI searchUpdated 1 min read

Also called thinking models · chain-of-thought models

What it means

A regular model starts writing its answer straight away. A reasoning model first works through the problem: it breaks it into steps, checks options and then answers. This takes longer but is better at hard questions.

OpenAI's o1, released in 2024, was the first widely used one. DeepSeek-R1 and Google's Gemini thinking models followed.

How it changes product answers

When a reasoning model compares products, it's more likely to:

  • Check requirements one by one, like budget, team size and integrations.
  • Run more searches to fill gaps.
  • Rule out options that don't clearly fit.

That rewards brands whose facts are easy to find and check. A vague page gives the model nothing to reason with.

Example

Asked for "a CRM under $30 per user that integrates with Outlook", a reasoning model checks each candidate's pricing and integrations pages. Two CRMs list Outlook clearly. One only mentions "email integrations". The model leaves the third out.

Frequently asked questions

Do I need to do anything different for reasoning models?

Mostly the same as for AI search in general, with more weight on clear, specific facts.

Are reasoning models used in ChatGPT by default?

ChatGPT picks a model depending on the question. Harder questions are more likely to get one.

Are reasoning models more accurate?

On complex questions, usually. They can still be wrong if the pages they read are wrong.

Go deeper

All posts →

Make AI search your next revenue channel

Start free trial