Reasoning refers to an AI model's ability to think logically and solve multi-step problems – rather than just reproducing patterns. Models with strong reasoning can solve math problems step by step, test assumptions, and draw complex chains of conclusions. Techniques like chain-of-thought and specialized "reasoning models" (e.g., OpenAI o1) aim at this.
In short
What is reasoning in AI?
Reasoning refers to an AI model's ability to think logically and solve multi-step problems – rather than just reproducing patterns. Models with strong reasoning can solve math step by step and draw complex conclusions. Techniques like chain-of-thought support this.
Using AI in your organisation?
We show you which applications are worth it for your organisation.