Automatic Prompt Engineer (APE)
Reasoning
Level 2 · Advanced reasoningGenerate multiple candidate phrasings of a prompt, evaluate them, and pick or combine the best.
Automatic Prompt Engineer asks the model to generate several different candidate phrasings for a prompt you need, evaluate each against stated criteria, and recommend the best one — or combine the strongest elements into a final version. It's distinct from Meta Prompting (which designs a strategy for solving a new task) and from the Prompt Optimizer (which improves a prompt you've already written): APE is specifically about generating and comparing multiple fresh phrasings before you've committed to any of them.
When to use it
Use it when you know what you want a prompt to accomplish but aren't sure how to phrase it most effectively.
When to avoid it
Avoid it once you already have a working prompt you just want to improve — that's the Prompt Optimizer's job.
Advantages
- Compares genuinely different phrasings instead of committing to the first one that comes to mind
- Makes the evaluation criteria explicit, not just a gut feeling about which phrasing "sounds better"
- Can surface a combination of elements from different candidates that's stronger than any single one
Disadvantages
- Adds a generation-and-evaluation step before you even start using the prompt
- Only as good as the stated evaluation criteria
Examples
I need a prompt that accomplishes: get concise, non-technical explanations of complex scientific concepts for a general audience. Generate 4 different candidate prompt phrasings for this task. Evaluate each against: clarity of instruction, likely consistency of output, ease of reuse. Recommend the best one, or combine the best elements into a final version.
Pairs well with
Try it
Fill in the placeholders below and see the compiled prompt update live.
What you want to achieve.
How should the alternatives be compared? (e.g. cost, impact, time to ship)
- 33
- Words
- 222
- Characters
- 0/2
- 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.