An AI model is a mathematical structure trained on data to perform a specific task, such as classifying images, translating text, or generating responses.
A model is the result of a training process. During training, the model adjusts its internal parameters to minimize errors on the training data. After training, the model can be used to make predictions or generate outputs on new data — this is called inference.
Models vary enormously in size, complexity, and capability. A small model might classify emails as spam or not spam. A large foundation model might handle many different language tasks.