What is a GPT?
A family of large language models developed by OpenAI that generate human-like text.
Stands for: Generative Pre-trained Transformer
Pronunciation: /dʒiː piː tiː/
GPT explained in plain English
GPT (Generative Pre-trained Transformer) is a series of AI language models created by OpenAI. They are "generative" because they create text, "pre-trained" because they learn from vast datasets before being fine-tuned, and "Transformer" because they use the Transformer architecture.\n\nEach GPT version has been larger and more capable than the last, with GPT-4 demonstrating strong reasoning, coding, and analysis abilities.
Analogy
GPT is like an apprentice writer who has read the entire internet, practiced writing every type of content imaginable, and then received specialised coaching for particular tasks.
Example
A marketing team uses GPT-4 via API to draft social media posts, analyse customer feedback, and generate product descriptions — all customised through prompt engineering.
How is GPT used?
GPT models power ChatGPT, Microsoft Copilot, and thousands of third-party applications via the OpenAI API. Developers use them for writing, coding, analysis, and customer service.
Common misconceptions about GPT
GPT is not a single model but a family. "GPT" is often used colloquially to mean ChatGPT, but they are related products, not the same thing.
History
GPT-1 (2018) proved the pre-training approach. GPT-2 (2019) showed scalability. GPT-3 (2020) demonstrated emergent abilities. GPT-4 (2023) added multimodal capabilities. ChatGPT (2022) made GPT accessible to the public.
Related terms
People also read
- LLM
A type of AI model trained on vast amounts of text to understand and generate human language.
- Transformer
The neural network architecture that revolutionised AI by enabling models to process entire sequences at once.
- generative AI
An emerging transformative field with no formal definition.
- RAG
A technique that combines AI language models with external knowledge retrieval for more accurate answers.
- Token
The basic unit of text that AI language models process, which may be a word, part of a word, or punctuation.
- AI slop
Output from a generative AI system that favors quantity over quality.
- BERT
A model architecture for text representation.
- BLEU
A metric between 0.
- BLEURT
A metric for evaluating machine translations from one language to another, particularly to and from English.
- Chain-of-Thought Prompting
Asking an AI to show its reasoning step by step before giving a final answer, which often improves accuracy on complex tasks.