# LLM (Large Language Model)

> A large language model (LLM) is a neural network trained on very large amounts of text to predict the next token in a sequence. That training lets it read, summarize, classify, translate and write text, and many current models also handle images and code. Businesses typically use LLMs through a provider's API, such as Anthropic's Claude or OpenAI's GPT models, or run open-weight models on their own infrastructure.

## Why it matters for your business

The LLM is the language and reasoning engine inside most AI automations, chatbots and agents, so the choice of model affects quality, speed, cost and data handling. Models differ in how well they follow instructions, use tools, handle long documents and decline unsafe requests. Usage is usually billed per token, so long prompts and high volumes drive running costs. An LLM on its own has no access to your data or systems and can state wrong facts confidently, which is why most business deployments pair it with retrieval, tool access and validation.

## Example

A freight broker uses an LLM to read inbound quote requests from email. The model extracts origin, destination, weight, equipment type and pickup date into a structured record, writes a short summary for the rep and flags requests missing key details. The rep still sets the price, but no longer retypes each request into the quoting system.

Related terms: https://www.lumeor.studio/glossary/token/, https://www.lumeor.studio/glossary/context-window/, https://www.lumeor.studio/glossary/inference/, https://www.lumeor.studio/glossary/fine-tuning/, https://www.lumeor.studio/glossary/hallucination/

---

Source: https://www.lumeor.studio/glossary/llm-large-language-model/