OpenAI is rolling GPT-6 out across ChatGPT, bringing its latest model generation and a new Intelligent UI system to a weekly audience of more than 1.2 billion people. The release reaches Plus, Pro, Business and Enterprise users globally first in the Chat tab, with Free and Go following the next day. Enterprise access remains subject to workplace admin settings. Paid tiers use GPT-6 Sol, while Free and Go use GPT-6 Luna, with both tuned for everyday conversation. The models behind Work and Codex are unchanged.
Intelligent UI lets ChatGPT build the format of an answer around the task instead of forcing every request into text. GPT-6 can combine prose with graphics, tappable buttons, forms, charts and other interactive elements, choosing between side-by-side comparisons, diagrams, maps or a plain answer. OpenAI says this can place a cooking timeline beside a recipe, plot road-trip stops, let learners adjust inputs, or create on-demand tools such as savings calculators, bill splitters and in-chat games.
The system runs on a library of native, streamable components and a compiler that processes an interface while the model generates it. That allows the response to appear progressively on web and mobile without waiting for the entire answer. Training focused on decisions about content, layout, visuals and when interactive elements are warranted. OpenAI acknowledges that the model's design judgment and the range of experiences it can create still need work.
GPT-6 can also start answering while it continues to think, interleaving reasoning with a series of useful partial responses intended to remain cohesive. In OpenAI's internal evaluation of high-value everyday agentic tasks, GPT-6 Extra High began answering as quickly as GPT-5.6 Medium while scoring above GPT-5.6 Extra High overall. For web-search questions, GPT-6 Instant started 44% sooner on average than GPT-5.6 Instant, and the company says the new generation more often identifies when to search and finds support for its answers.
Safety work builds on Astra advances, with stronger resistance to multi-turn attempts to bypass safeguards and updated protections around cyberattacks, biological threats and violence. GPT-6 is also trained to use conversational context when risk is not visible from one prompt, avoid needless refusals, and state more clearly when it lacks the information or tools required.
The launch expands a model family introduced to paying customers last month and frames Intelligent UI as a move toward software that adapts to a person's goal. OpenAI says its longer-term aim is for people to describe what they want to accomplish and work inside an experience created around that task, advancing its mission to bring the benefits of artificial general intelligence safely to humanity.
Sources and related context #
- GPT-6 Sol and GPT-6 Luna October system card : Supports: The system card supports the reported safety advances and distinguishes the new Chat models from the September versions remaining in Work and Codex.
- OpenAI's model guide for the GPT-6 family : Related context: This earlier guide explains the GPT-6 family used for coding and API workflows, providing background on the models outside this Chat-focused rollout.