Fine-tuning — Fine-tuning is additional training that adapts a pretrained model to a specific task, style, or field. It uses a much smaller, focused dataset. A company might fine-tune a model on its own support conversations, for example. Fine-tuning changes the model’s weights, unlike prompting.

Added October 21, 2026