AI glossary
Large language model (LLM)
An AI model trained on huge amounts of text to predict the next token, which lets it write, summarise, translate, answer questions and write code.
An LLM reads text as tokens and, at each step, predicts which token is most likely to come next. Repeating that step produces whole answers. Because it learned from a vast amount of writing, it picks up grammar, facts, reasoning patterns and coding styles along the way.
That is also why LLMs can be confidently wrong: they generate what is plausible, not what is verified. Claude, GPT and Gemini are families of LLMs.
Example: When you ask “What is the capital of France?”, the model produces “Paris” because that continuation is overwhelmingly likely given what it learned.
In practice
- Treat its answers as a good first draft: check facts and figures.
- The more context and examples you give it, the better it answers.
- Each family has models of several sizes; compare them in our model guide.
The full explanation is in how language models work.
