The basics
In AI, a model is the trained program that has learned patterns from data and can now make predictions or generate output for new inputs.
A model is what you get at the end of training. Before training it is a blank structure full of random numbers. After training, those numbers have been tuned so that inputs produce useful outputs: a photo comes in and 'golden retriever' comes out, or a question comes in and an answer comes out.
Think of a model as a very complicated function, or as a trained skill. Its 'knowledge' is not stored as sentences you could read. It lives in millions or billions of numbers called weights. That is why you cannot simply open a model and edit a fact; you retrain or fine-tune it instead.
People say 'the model' for many sizes of thing. A spam model might be a few thousand numbers. A large language model can be hundreds of billions. Names like GPT, Claude, Gemini, or Llama refer to specific models or model families, as of 2026.
When you point your phone camera at a restaurant menu abroad and the words turn into English on screen, a translation model is doing the work.