What is an LLM?

LLM stands for large language model: the technology behind ChatGPT, Claude, Mistral, and most other AI services you chat with. An LLM has been trained on huge amounts of text and learned to predict which word is most likely to come next. From that seemingly simple ability, everything else grows: summarising, translating, reasoning, and drafting.

"Large" refers to the number of adjustable values in the model, often hundreds of billions. Larger models usually reason better but cost more per answer, which is why you choose the model based on the task. A simple rephrasing doesn’t need a giant model, but a tough analysis does. The glossary has a fuller explanation under the entry for språkmodell, the Swedish term for the same thing.

In EuroWork, the model choice is a visible part of the service. The quality selector typically shows the price per answer before you send the question. The standard models are European, and every answer tells you which model was used and where the processing happened.