GPT (“generative pre-trained transformer”) is OpenAI’s line of language models behind ChatGPT and the OpenAI API. After GPT-4 (2023), GPT-4o (2024) and the o-series reasoning models, GPT-5 (August 2025) merged fast answers and reasoning into one family. Updates then arrived every few months: GPT-5.1 to GPT-5.6.
The GPT-6 generation launched in September 2026 with planet-named tiers: Astra (most capable), Sol (balanced), Terra (fast, with visual reasoning and computer use) and Luna (cheapest). GPT-6.1 Sol and Terra followed in September and October. Every current model takes text and images and has a context window of about 1.05M tokens.
OpenAI retires models quickly. GPT-5, o3, GPT-4.1 and GPT-4o all have shutdown dates between December 2026 and January 2027, so check the lifecycle table before you pin a model ID in production.
