Term · AI
LLM (large language model)
An LLM, or large language model, is the engine behind services such as ChatGPT, Copilot and Claude. It writes text by repeatedly predicting the next word.
In short
An LLM (large language model) is an AI model that has been trained on a very large amount of text. It produces text by repeatedly predicting which word fits best after the words before it. That makes it write fluently, but it does not check whether something is correct.
What is an LLM?
An LLM is a large language model: an AI model that can read and write text. LLM stands for large language model. It has been trained on enormous amounts of text and has learned the patterns of language from it.
You rarely talk to a language model directly. You use a service that runs on it. ChatGPT from OpenAI, Microsoft Copilot and Claude from Anthropic are examples. The language model is the engine. The service is the dashboard you use to operate that engine.
How does a language model work?
A language model splits text into small pieces, tokens. These are whole words, parts of words or punctuation marks. At each step, the model predicts which piece fits best after the text before it. That way it builds up an answer word by word.
That works well for writing, summarising and translating. But the model does not look up what is true. It chooses what sounds likely. That is why it can make up something that seems convincing: a hallucination.
A language model also only knows what was in its training. It does not know your own quotes or manuals. With RAG you have the model search your own documents first.
Example in an SME
A made-up example
A wholesaler uses a language model to prepare replies to customer emails. The company works with a business account with agreements about the data. The employee gets a draft and checks prices and delivery times in the company's own system before he replies.
What should you look out for?
- Choose a business version. With it, you make agreements about your data, such as a data processing agreement. According to OpenAI and Microsoft, data from their business versions is not used by default to train language models.
- Check facts. You look up names, amounts, dates and legal articles yourself.
- Choose what fits. Which model or service fits depends on your work and the software you already have. Read more in Which AI suits your business?
A language model can also be part of your own software. Think of an assistant that answers questions from your manuals. I build such applications as custom work, in the systems you already use. Read more about AI in your business.
Frequently asked questions
What is the difference between an LLM and ChatGPT?
An LLM is the model itself. ChatGPT is a service from OpenAI that runs on such models, with a chat screen and extra features around it.
Does a language model know what is true?
No. It predicts which text fits. Often that is correct, sometimes not. That is why you check important facts yourself.
Does a language model learn from what I enter?
That depends on the service and your settings. With the personal version of ChatGPT, OpenAI may use your conversations to train models, unless you turn that off. With ChatGPT Business and Enterprise, according to OpenAI, that does not happen by default.
Sources: OpenAI: How ChatGPT and our foundation models are developed, OpenAI: how your data is used to improve model performance, Microsoft Learn: Data, Privacy, and Security for Microsoft Copilot and Dutch Data Protection Authority (Autoriteit Persoonsgegevens): Guidance on generative AI and the GDPR (in Dutch). Consulted on 7 October 2026.
Written by Bart van Engelenhoven Founder of Uniekaiconsultancy in Ede, the Netherlands. Bart analyses how SMEs work and builds software, integrations and AI applications himself. Last updated: 7 October 2026
Read more
Term
Generative AI
Generative AI creates new text, images or summaries based on an instruction. What it is, how it works and what to watch for as a business.
Read moreTerm
RAG (retrieval-augmented generation)
RAG lets a language model search your own documents first and then answer, with a reference to the source. How it works and what to look out for.
Read moreTerm
AI hallucination
An AI hallucination is an answer that sounds convincing but is wrong. Why language models do this and how to check an AI answer.
Read moreNext step
Curious what could work for you?
- The first conversation is free and without obligation
- By phone or at your premises
- You speak directly with me
Bart van Engelenhoven You tell me what takes up your time, I tell you honestly whether I can help.