Zero-Shot Prompting

Zero-shot prompting means asking the model to perform a task without giving it any examples — just a direct instruction like 'Summarize this article in three bullet points.' Modern models are strong zero-shot learners because their training exposed them to countless variations of common tasks.

It's the fastest, simplest way to prompt and works well for clear, common tasks. When the task is unusual, highly structured, or the model keeps misunderstanding the desired format, that's the signal to add examples (few-shot) instead.

Zero-shot is the natural baseline: try it first, and only add complexity when it falls short.