OpenAI Prompt Engineering Guide logo

OpenAI Prompt Engineering Guide

Open Source

OpenAI's official guide to crafting effective prompts and best practices for GPT models.

Visit website

Pricing

Free / Open source

Type

Manual

// VERDICT

Use the OpenAI prompt-engineering guide as the authoritative reference for prompting OpenAI models. It's documentation, not a tool - pair it with a prompt-testing tool (PromptFoo) and an eval framework to verify and measure prompts.

Best for

OpenAI's official prompt-engineering guidance - documentation and best practices for prompting OpenAI models, not a tool you install.

Avoid when

You want an executable tool, model-agnostic guidance, or automated prompt testing rather than reference material.

CI/CD fit

Reference guidance - pair with a prompt-testing tool for CI

Team fit

Anyone prompting OpenAI models · Prompt engineers · QA writing eval prompts

Setup

Easy

Maintenance

Low

Learning

Beginner

Licence

Free / Open source

// BEST FOR

  • Learning effective prompting for OpenAI models
  • Authoritative, vendor-maintained best practices
  • Patterns for instructions, structure and examples
  • Improving prompts before testing them
  • A free reference for teams on OpenAI
  • Grounding prompt reviews in documented guidance

// AVOID WHEN

  • You want an executable tool, not documentation
  • Model-agnostic guidance is needed
  • You want automated prompt testing (PromptFoo)
  • You need measurable evals, not advice
  • A hosted platform is the goal
  • You're not using OpenAI models

// QUICK START

Read OpenAI's prompt-engineering guide and apply its techniques (clear
instructions, structure, examples) -> verify prompts with a prompt-testing tool
and measure quality with an eval framework.

// ALTERNATIVES TO CONSIDER

ToolChoose it when
Claude Prompt Engineering GuideYou're prompting Claude and want Anthropic's guidance.
PromptFooYou want to actually test and compare prompts.
OpenAI EvalsYou want to benchmark model behaviour with evals.

// FEATURES

  • Six core strategies illustrated with concrete prompt examples
  • Guidance on system, user, and developer message roles
  • Patterns for structured output, function calling, and JSON mode
  • Cookbook repository with worked notebooks
  • Playground for iterating on prompts against live models

// PROS

  • Authoritative source from the GPT model provider
  • Strong coverage of structured output and tool use
  • Freely accessible, no paywall on documentation
  • Cookbook examples are runnable and well-maintained

// CONS

  • Tuned to GPT models — patterns may differ for other providers
  • Less prescriptive on evaluation than dedicated eval tools
  • Spread across docs, cookbook, and platform — discoverability suffers

// EXAMPLE QA WORKFLOW

  1. Read the OpenAI prompt-engineering guide

  2. Apply techniques to your prompts

  3. Add structure and examples

  4. Verify prompts with a testing tool

  5. Measure quality with an eval framework

  6. Iterate as models/needs change

// RELATED QA.CODES RESOURCES

Cheat sheets