OpenAI announced the release of GPT-6 models for ChatGPT, bringing the next-generation AI to the more than 1.2 billion people who use the service each week. The headline feature is Intelligent UI, a capability that lets ChatGPT compose answers that combine text with graphics, tappable buttons, forms, charts, and interactive experiences tailored to the user's question. For example, when asked to break down the design of a seven‑speed bicycle, the model can generate an interactive diagram where users explore each subsystem—frame, drivetrain, brakes, and cockpit—and see how they work together.
Intelligent UI is powered by GPT‑6 Instant, which enables the model to begin delivering an answer while it continues to think, interleaving reasoning with response generation. OpenAI built a library of native, streamable UI components and a compiler that processes the interface as the model generates it, allowing parts of the interface to appear progressively without waiting for the full response. Training involved evaluating the generated interfaces for clarity, usefulness, and completeness, teaching the model to decide when interactivity aids understanding versus when a simple text answer suffices, and to organize information clearly across layouts.
Alongside the intelligence gains, GPT‑6 incorporates safety advances from its Astra lineage, including updated training to resist adversarial attempts to bypass safeguards and better recognize its own limitations. The model is also trained to avoid unnecessary refusals to harmless requests by using conversation history to contextualize risks. Rollout starts today for ChatGPT Plus, Pro, Business, and Enterprise subscribers using GPT‑6 Sol, and expands tomorrow to Free and Go tiers using GPT‑6 Luna, with enterprise access subject to workspace admin settings.
This image illustrates the GPT-6 model family discussed in the article, showing the different variants such as GPT‑6 Sol and GPT‑6 Luna.