AI glossary · Models
Language model
In one sentence
A language model is an AI system that calculates which word, or token, is most likely to come next in a piece of text. By repeating that prediction it can complete sentences, answer questions, summarise or translate.
Also known as: LM · natural language model · text model
What is a language model?
The idea is old and simple: read enough text and you can guess how a sentence carries on. After “I can confirm your appointment for…” you expect a day of the week, not a recipe. A language model turns that intuition into probabilities.
Put it to work
AI agents that look after your customers
From €49/mo + VAT
Voice, WhatsApp and email with the phone line included: agent, numbers and CRM in one fee.
See plans and pricingEarly models just counted which words appeared together. Today’s are neural networks trained on huge collections of text, and nearly all of them are built on the transformer architecture. Depending on size, people talk about LLMs, the large ones, or small language models, which are cheaper to run.
The key point: a language model produces text that sounds right. Whether it is right depends on the data you give it and how it’s used.
Related to this term
UTIO agents
- Available 24/7 in 26 languages
- Set it up yourself for €0: built from your website
- 14-day free trial, no card needed
From €49/mo + VAT
See plans and pricingHow it works
A language model has two distinct phases:
- Training. It reads billions of sentences and adjusts its internal parameters to get better at guessing the next word. This is expensive and happens once, or only occasionally.
- Use. It takes your text, splits it into tokens, calculates the probability of every possible next token and picks one. It appends it and repeats, building the reply piece by piece.
Everything else is built on top of that. With the right instructions, the same model can sort emails, pull the details out of an order or draft a reply. With fine-tuning, it can be specialised in a particular style or task.
Example
A car repair garage in Bristol gets messages like: “hiya, my car’s making a weird noise when I brake, any chance you could look at it Thursday?” A language model can pull out three useful facts: the reason (a possible brake fault), the intent (booking a slot) and the preferred day (Thursday).
With those, the system can offer free slots and log the booking. Nobody had to write a rule for every possible way of asking for an appointment: the model copes with the variety of everyday language.
How it affects your business
Behind every assistant, every search engine that answers in full sentences and every AI agent, there’s a language model. Knowing what it is helps you hand it what it does well (drafting, summarising, classifying, conversing) and not leave it alone with what it can’t guarantee, such as an exact price or a legal deadline.
At UTIO, the language model is what lets the WhatsApp agent and the voice agent understand messages however they’re worded. What they say about your business isn’t made up on the spot: it comes from your knowledge base, through RAG.
Frequently asked questions
Is a language model the same as an LLM?
An LLM is a large language model. Every LLM is a language model, but not every language model is large: there are small language models (SLMs) and older statistical models, like the ones that suggested the next word on your phone keyboard.
Does a language model understand what it says?
It depends on what you mean by understand. It represents the meaning of words well enough to summarise, classify or answer accurately, but it has no intent and doesn't check on its own whether what it says is true. That's why it pays to give it sources and review anything important.
Related terms
- LLM
An LLM (large language model) is an AI model trained on vast amounts of text to predict the next word. That single skill lets it draft, summarise, translate, classify and hold a conversation.
- Small language model (SLM)
A small language model (SLM) is a language model with far fewer parameters than an LLM. It knows less about everything, but it replies faster, costs less and can run on a modest server, a laptop or even a phone.
- Natural language processing (NLP)
Natural language processing (NLP) is the branch of artificial intelligence that lets machines work with human language: read, analyse, classify, translate and produce text or speech the way people use it.
- Token
A token is the chunk of text a language model works with: a word, part of a word or a punctuation mark. Models read, count and bill in tokens, not in words or characters.
- Transformer
A transformer is a neural network architecture that processes every word in a text at once and uses an attention mechanism to work out which parts relate to each other. It's the foundation of almost every language model in use today.
Updated on