GPT originally stood for Generative Pre-trained Transformer, the name OpenAI gave the model architecture introduced in its 2018 paper "Improving Language Understanding by Generative Pre-Training." The architecture combined three ideas: the transformer (Vaswani et al., 2017), self-supervised pre-training on large text corpora, and fine-tuning for specific downstream tasks.
The GPT name has since become more specific (referring to OpenAI's commercial model releases: GPT-2, GPT-3, GPT-3.5, GPT-4, GPT-5, GPT-5.4) and more diffuse (used colloquially to refer to any large language model). CrashBytes coverage generally uses "GPT" to refer specifically to OpenAI's models, and "LLM" for the broader category.
The GPT lineage has been the commercial and cultural catalyst for the current AI wave. The public launch of ChatGPT in November 2022 — built on GPT-3.5 — is widely credited as the moment the modern AI industry became a mainstream category.