What the letters mean
GPT is OpenAI’s series of large language models. It stands for Generative Pre-trained Transformer, and the name states the three key elements directly: it generates text, it is pre-trained on an enormous corpus, and it uses the Transformer architecture.
GPT and ChatGPT
These get confused constantly. GPT is the model; ChatGPT is the service that makes it usable through conversation. GPT is the engine, ChatGPT is the finished car. When you see a numbered name such as GPT-5 in the news, it is a generation of the engine.
How the series developed
The series began in 2018 and grew larger and more capable with each generation. GPT-3 in 2020 drew researchers’ attention as a model that wrote startlingly natural prose. In 2022, ChatGPT packaged GPT for conversation and set off the global generative AI boom. Development continued with multimodal capability covering images and audio, and with reasoning models that think before answering.
From product name toward generic term
GPT is properly OpenAI’s model name, but its influence has led to the letters appearing in other custom AI features — such as GPTs — and sometimes used loosely for LLMs in general. Being able to tell whether a piece of writing means OpenAI’s model or the category as a whole makes it much easier to read accurately.
Specific model names and capabilities change often, so check OpenAI’s own announcements for current information. Holding on to GPT as the name of the model series inside ChatGPT is enough to follow launch coverage.