OpenAI Rolls Out GPT-6 To All ChatGPT Users, Adds New 'Intelligent UI' That Offers Interactive Answers

OpenAI began rolling out GPT-6 in ChatGPT, extending the model's access beyond paying customers to the more than 1.2 billion people the company says use the chatbot every week. The release comes a month after the first GPT-6 models reached paid users. Its headline addition is Intelligent UI, a feature that lets ChatGPT answer with visuals and interactive elements instead of text alone.

What Is Intelligent UI?

OpenAI says it trained GPT-6 to decide how to present each answer, mixing text, graphics, tappable buttons, forms and charts depending on the question. A comparison might appear side by side, while an explanation could take the form of an interactive diagram. A plain text answer is still an option when that serves best. OpenAI's examples include a road trip plotted on a map with notes on detours, a Sunday roast plan with a guest-count slider that recalculates shopping quantities, and explainers where users change an input to see the result. Users can also ask ChatGPT to build a small tool on the spot, such as a bill splitter, a retirement savings calculator or a game.

Technically, OpenAI built a library of native, streamable interface components and a compiler that renders the interface progressively as the model generates it. The company acknowledges there is 'still work ahead' on the model's design judgment.

Answers that start before thinking ends

GPT-6 can also begin answering while it is still reasoning, building a response across several partial replies. OpenAI says that in an internal evaluation of everyday agentic tasks, GPT-6 Extra High started answering in the same time as GPT-5.6 Medium while scoring higher than GPT-5.6 Extra High. For questions needing web search, GPT-6 Instant begins answering 44 percent sooner on average than GPT-5.6 Instant. These are OpenAI's internal evaluations and have not been independently verified.

Who gets what, and when

The rollout has begun for Plus, Pro, Business and Enterprise users in the Chat tab, and expands to Free and Go users a day after. Enterprise access depends on workplace admin settings. Paid tiers run on GPT-6 Sol, while Free and Go users get GPT-6 Luna. The models behind ChatGPT's Work and Codex experiences are unchanged in this release.

Safety claims

OpenAI says GPT-6 builds on safety advances from its Astra model and resists attempts to bypass safety training better than GPT-5.6 Sol, particularly multi-turn attacks. It also says the model is better at flagging what it cannot do and at avoiding unnecessary refusals. The company points to a system card for detail.

The Luna model now serving free users drew attention on the same day. Anthropic's launch of Claude Haiku 5.5 benchmarked the new small model against GPT-6 Luna, claiming leads on several tests. Both sets of figures come from the companies themselves.