Also called: LLM · Last checked: 7 November 2026
In short: A large language model (LLM) is an AI system trained on huge amounts of text to predict the next word. That one skill lets it write, summarise, translate, answer questions and draft code.
Table of Contents
How it works
- Training. The model reads billions of pages of text and learns which words and ideas tend to follow each other. What it learns is stored as numbers called parameters.
- Tuning. People rate its answers, so it learns to follow instructions and refuse harmful requests.
- Use. When you type a prompt, it writes the reply one token (a word or part of a word) at a time.
Where you’ll meet it
ChatGPT, Claude, Gemini and Microsoft Copilot all run on large language models. Meta’s Llama models are open-weight, so a company can run them on its own servers.
What it means for your business
- Language work gets cheaper. First drafts, summaries and routine customer replies take seconds.
- You rent the model. Most firms pay a monthly fee or pay for usage, rather than build their own.
- Your own data sets you apart. Rivals can buy the same model. The gain comes from linking it to your documents and workflows.
Watch out for
- Made-up answers. A model can state false facts with confidence. Check anything that matters.
- Out-of-date knowledge. It knows only its training data unless it can search the web or your files.
- Leaked data. Staff pasting customer details into free tools can break privacy rules.
Example
Klarna. In February 2024 Klarna said its OpenAI-powered assistant handled 2.3 million chats in its first month, two-thirds of all customer service chats and the work of about 700 full-time agents. In May 2025 its chief executive said the push had lowered quality, and Klarna began a pilot to bring back human agents so customers could always reach a person.
Large language model vs foundation model
| Large language model | Foundation model | |
|---|---|---|
| What it is | A model trained mainly on text | Any large model trained on broad data and adapted to many tasks |
| Works with | Words and code | Text, images, sound, video or a mix |
| How they relate | One kind of foundation model | The wider family, which includes LLMs and image models |
Frequently asked questions
Is ChatGPT a large language model?
No. ChatGPT is an app. It runs on large language models made by OpenAI and adds features such as web search, memory and file uploads.
Do large language models understand what they write?
Not in the way people do. They predict likely words from patterns in their training data, which is why fluent text can still be wrong.
How can a small business use a large language model safely?
Use a business plan that does not train on your data, keep personal data out of prompts, and have a person check anything that goes to customers.