Zero-Shot Prompting
Asking the model to do a task with no examples, relying purely on its instructions and training.
It is the default way most people prompt. Knowing when to switch to few-shot is the skill.
Zero-Shot Prompting in practice
Most everyday prompts are zero-shot. When the answer comes back in the wrong shape or tone, the usual fix is to add one to five examples and turn it into few-shot prompting. Our prompting guide shows where examples fit in a full prompt.
Frequently asked questions
What is zero-shot prompting?
Zero-shot prompting is asking the model to do a task with no examples at all, relying purely on your instructions and its training. It is how most people prompt by default: you just ask.
Is zero-shot prompting good enough for most tasks?
Often yes, especially for common requests the model handles well without hand-holding. The real skill is recognizing when zero-shot falls short and switching to a few examples to lock in the result.
When does zero-shot stop being the right choice?
When you need a precise format, a particular style, or consistent results that the model keeps getting slightly wrong. That is the signal to add a couple of examples and move to few-shot prompting.
New to all this? Start with what an AI agent really is, browse the full glossary, or explore the learning hub.
← Back to the glossary