Large language model
A large language model is a system trained on very large amounts of text to predict what text should come next, which is what lets it write, summarise, classify and answer — and also why it can produce something fluent and wrong.
Also written: LLMArabic: النموذج اللغوي الكبير
What that definition implies
Fluency and accuracy are separate properties. The model is optimised for text that reads correctly, and text that reads correctly is not the same as text that is correct. Every practical decision about deploying one follows from that gap.
What they are good and bad at in a business
Good at transforming text you already have — summarising, extracting, reformatting, drafting. Poor at being a source of fact about your business, which is what retrieval is for. Poor at arithmetic and at anything needing a guaranteed answer.
The question to ask before using one
Not "can it do this" — it usually can, once. Ask what happens the time it is wrong, who notices, and what that costs. If nobody would notice, the task is not ready for a model.
Questions people ask
Almost never. Most business value comes from connecting a general model to your own data and constraining what it is allowed to do, which is faster, cheaper and easier to change.
The one that passes your evaluation set. Model rankings move monthly; a test built from your real cases outlives them.
Related serviceAI for Business