OpenAI is rolling GPT-6 with Intelligent UI into ChatGPT globally, expanding the release from paid tiers on October 7 to Free and Go users starting October 8.
The model update changes more than the name in the selector. OpenAI says ChatGPT can now compose answers from text, visuals and native interactive components such as charts, forms, tappable buttons, diagrams and small tools, choosing the format based on the request. The rollout is staged, so "starting October 8" does not mean every eligible account flips over at the same instant.
Paid plans get Sol while Free and Go get Luna
For ChatGPT Plus, Pro, Business and Enterprise, the new everyday Chat model is GPT-6 Sol. Free and Go users receive GPT-6 Luna, an optimized version aimed at the same everyday-chat experience.
OpenAI says Plus, Pro, Business and Enterprise began receiving GPT-6 globally on October 7, while Free and Go rollout starts October 8. Enterprise availability can also depend on workplace admin settings.
The company calls the presentation layer Intelligent UI. Instead of treating every response as a stream of paragraphs, GPT-6 can use a library of native components and a compiler that renders interface elements progressively while the model continues generating. OpenAI's examples include comparison layouts, interactive diagrams, calculators, bill splitters and small games.
The Verge's October 7 coverage independently confirms the broad product shape: more visual and interactive responses, a split between Sol on higher tiers and Luna on Free/Go, and a global rollout that begins with paid users before expanding.
The answer can start before all the work is finished
OpenAI also says GPT-6 can begin presenting useful parts of an answer while it continues thinking or using tools, then add findings as they become available. That is meant to reduce the dead time before a complex answer becomes readable.
There is an important product boundary in the announcement. This release changes the Chat experience. The models powering Work and Codex are not changing as part of it. That prevents the GPT-6 Chat rollout from being read as a simultaneous model replacement across every OpenAI product.
OpenAI says more than 1.2 billion people use ChatGPT each week and reports improvements in its internal safety and honesty evaluations. Those are company figures and company evaluations, not independent measurements, so they belong with that attribution attached.
For users, the most visible change is likely to be the interface rather than a benchmark score. A question can now produce a normal text answer, a visual explanation, or an interactive control when the model judges that format more useful.
That also means rollout quality will be judged on restraint. A clever diagram is useful when it clarifies something; a surprise button farm is not. OpenAI's own pitch says text remains available when text is the best answer.
As of October 8, the direction is clear even if account-by-account availability is still catching up: GPT-6 is becoming the default ChatGPT generation for a much wider audience, and Intelligent UI is turning the response surface itself into part of what the model produces.
Reporting notes