Large language model (LLM)
A large language model is an AI model trained on large volumes of text that processes and generates language.
At their core, LLMs predict the most probable next piece of text. This simple principle yields broad applicability: summarising, classifying, translating, extracting data from unstructured text and holding conversations.
Well-known model families come from OpenAI, Anthropic, Google, Meta and Mistral, among others. They differ in capability, cost and operating model — from purely cloud-based services to models that can be run in your own data centre.
For enterprise use, two questions are central: where the data is processed, and where the model gets its knowledge. In practice the second question almost always leads to retrieval-augmented generation.
AI in mid-sized companies: use cases that actually pay off
Less hype, more arithmetic: which AI applications deliver measurable value in mid-sized companies, why projects fail and what a sensible start looks like.
Read the article →Related service
Let's talk about your project.
Free initial consultation, 30–45 minutes, remote. An honest assessment — even if the answer is that you don't actually need it.