Large Language Model (LLM)
A large language model is a system trained on very large amounts of text to predict likely continuations. That single ability underlies writing, summarising, classifying, extracting and answering — it is not a database and does not look facts up unless connected to one.
Because a model predicts plausible text rather than retrieving stored facts, it will produce a fluent, confident, wrong answer as readily as a right one. This is not a bug being fixed; it's what the mechanism does. Reliability comes from the system built around the model — grounding it in real documents, constraining it to a set of tools, checking its output.
Practically, models differ on three axes that matter more than benchmark scores: cost per token, latency, and how well they follow instructions under pressure. A cheaper, faster model that reliably returns valid JSON often beats a smarter one that occasionally decides to explain itself instead.
Why it matters
The question is never "which model is best" but "what happens when it's wrong." A business process that can't tolerate a confident error needs a verification step, not a better model.
Related terms
More in AI Automation & Agents
Need this built rather than explained?
We publish every price we charge, and you get a quote in writing before anything starts.