# Large language model

> The kind of AI behind chat tools: trained on huge amounts of text, it generates answers by predicting the next word, over and over.

**Category:** Core concept, Under the hood

**A large language model, or LLM, is the engine behind tools like Claude.** It is trained on a huge sweep of text, then generates answers by **predicting the next word over and over**, which is why it is so fluent.

"Large" is literal: these models have billions of internal settings tuned during [training](/glossary#training). Bigger, better-trained models can usually handle harder tasks. It is the technical name for the [model](/glossary#model) you actually chat with.

## At work

Every time Claude drafts an email or summarises a report, an LLM is doing the work: predicting the next word until the answer is complete.

## Often confused with

AI in general. An LLM is one kind of AI model, the text-generating kind, not a synonym for all AI.

## Related terms

- Model
- Generative AI
- Token

## See it in a lesson

[How generative AI works](/en/learn/foundations/how-generative-ai-works)

## Further reading

- [What are large language models? (IBM)](https://www.ibm.com/think/topics/large-language-models): A plain explainer of how LLMs are built and used.
- [Large language model (Wikipedia)](https://en.wikipedia.org/wiki/Large_language_model): A fuller, well-sourced reference.
- [Claude models overview (Anthropic docs)](https://docs.claude.com/en/docs/about-claude/models/overview): The specific LLMs Claude offers.

---

Last updated: 24/07/2026  
Source: Rollo Academy glossary