Overview

What is GPT?

GPT (Generative Pre-trained Transformer) is a family of large language models developed by OpenAI.
These models excel at generating natural language, answering questions, and performing reasoning tasks, thanks to training on vast amounts of text data.

GPT models are widely used for chatbots, content creation, coding assistance, data analysis, and more.

Key features

Natural language understanding and generation
Strong reasoning abilities
Multi-turn conversation handling
Multilingual support
API and tool integration (functions, retrieval-augmented generation)

Strengths

Versatile across many domains (technical, creative, analytical)
Excellent for structured output (tables, JSON, formatted reports)
Strong code generation and debugging abilities
Works well with role-based prompting

Limitations

May produce plausible but incorrect information (hallucinations)
Can be sensitive to prompt phrasing
Default model may lack up-to-date real-world knowledge (unless connected to browsing or plugins)

Best use cases

Conversational AI for customer support
Content creation (articles, marketing copy, scripts)
Code generation in multiple programming languages
Data analysis and summarization
Educational tools and tutoring

Prompting tips for GPT

Use clear, specific instructions for better accuracy
Set output constraints (word count, format, style)
For complex tasks, break them down into steps
Use system messages (if available) to define role and behavior
Include examples for few-shot prompting when needed

Model compatibility in this documentation

All prompts marked as GPT-compatible in this collection have been tested primarily with GPT-4 and GPT-4 Turbo.
Many also work well with GPT-3.5, with slight adaptations.