GPT — Generative Pre-trained Transformer — is the family of large language models built by OpenAI. GPT-2 (2019) demonstrated the promise of scaling. GPT-3 (2020) proved commercial viability. GPT-4 (2023) hit widespread enterprise adoption. GPT-5 (2025) became the default general-purpose LLM for many production applications.
GPT is a transformer-architecture model trained on a massive text corpus with reinforcement learning from human feedback (RLHF) for alignment. It powers ChatGPT (OpenAI's consumer product) and is available via OpenAI's API for developers.
GPT is the single most widely-used LLM in production in 2026 — though Claude, Gemini, and open-source Llama models have all gained meaningful share for specific use cases.