Zero-shot prompting
Fundamentals
Level 1 · FoundationsAsk the model to complete a task directly, with no examples at all.
Zero-shot prompting relies entirely on the model's pretrained knowledge: you describe the task and the desired output, and the model attempts it without seeing any worked examples first. It works because large language models have already seen countless similar tasks during training, so a clear, well-specified instruction is often enough.
When to use it
Use it for common, well-understood tasks (summarizing, translating, classifying, rewriting) where the desired output is easy to describe in words and the model has almost certainly seen similar tasks during training.
When to avoid it
Avoid it for tasks with an unusual or highly specific output format, domain-specific conventions, or edge cases that are hard to describe precisely — in those cases the model needs to see an example, not just a description. A zero-shot prompt should also tell the model what to do when information is missing: explicitly instructing it to state what's missing instead of inventing information meaningfully reduces the model's tendency to fabricate details it doesn't actually have.
Advantages
- Fastest technique to write — no examples to prepare
- Uses the fewest tokens, which keeps costs and latency down
- Scales well: the same instruction works across many similar inputs
Disadvantages
- Less reliable for tasks with ambiguous or unusual output formats
- No way to show subtle edge cases or exceptions
- Quality depends heavily on how precisely the instruction is worded
Examples
Classify the sentiment of this review as positive, negative, or neutral. Review: "The battery life is incredible, but the app crashes constantly." Output format: a single word.
Pairs well with
Try it
Fill in the placeholders below and see the compiled prompt update live.
What you want to achieve.
The material to work from.
The structure of the output.
- 27
- Words
- 199
- Characters
- 0/3
- Placeholders filled
- 1 min
- Reading time
- Completeness0
- Clarity0
- Context0
- Specificity0
- Constraints0
- Format0
- Output0
- Constraints note0
- Success criteria0
- Scope focus100
- Verification step0
- Negative example0
- Confidence0
- Placeholder syntax100
- Markdown structure0
- Fill in the remaining placeholders.
- Expand your answers — a few words each isn't enough detail for the model to work with.
- Add background context so the model understands the situation.