Model Lifecycle
Fine-tuning
Additional training on top of a pretrained model to specialize its behavior — for a task, a style, or a company’s own data.
Where pretraining teaches a model language and general knowledge, fine-tuning shapes what it does with that knowledge: following instructions, refusing unsafe requests, matching a brand's tone, or getting good at a narrow task like reviewing legal documents.
Fine-tuning is far cheaper than pretraining because it starts from an already-capable model and needs a much smaller, more targeted dataset — often thousands rather than trillions of examples.