TL;DR
OpenAI rolled GPT-6 into ChatGPT on October 7 with a feature it calls Intelligent UI: answers can now include tappable buttons, forms, charts, and small working tools like a dinner bill splitter, chosen to fit the question. Paid tiers (Plus, Pro, Business, Enterprise) got GPT-6 Sol first; Free and Go users started receiving GPT-6 Luna today, October 8.
What changed
On October 7, OpenAI announced GPT-6 for ChatGPT with what it calls Intelligent UI — the model now composes replies from text, visuals, and interactive elements, picking the format based on the question. A comparison can appear side by side. An explanation can turn into an interactive diagram. When plain text is the most useful answer, ChatGPT still answers in plain text.
The rollout started globally in the Chat tab for Plus, Pro, Business, and Enterprise subscribers on October 7, and expanded to the Free and Go tiers today, October 8. OpenAI says more than 1.2 billion people use ChatGPT each week, so this is a product-surface change at enormous scale.
What it can build for you
OpenAI's examples read like a list of small apps you'd never bother installing: a recipe with the cooking timeline shown next to it, road trip stops placed on a map, an interactive explanation where you change an input and see what happens (their demos covered the central limit theorem and the Monty Hall problem), a retirement savings calculator, a dinner bill splitter, even a game you play inside the conversation.
Sam Altman's framing on X was one sentence: "ChatGPT can now generate a custom UI for you." The point is practical — instead of describing what a calculator should look like, you get one you can actually use. Users can also dial the visuals back down when they want: TechCrunch, which attended the briefing, reported a setting to reduce the amount of visual material in responses.
How it works under the hood
Behind the feature is a library of native, streamable components and a compiler that processes the interface as the model generates it. That means the screen appears piece by piece instead of waiting for the whole response to finish — a genuine difference from the old pattern of dumping a finished block of text.
GPT-6 also changed the answering rhythm. Instead of waiting until the model has completed its entire reasoning process, it can begin answering while continuing to work through harder parts of a request, adding findings to the response without another prompt. OpenAI's company-reported number: GPT-6 Instant begins responding 44% sooner on average than GPT-5.6 Instant for questions requiring web searches, based on internal testing.
How to try it, and the catches
Check the ChatGPT model picker: paid tiers use GPT-6 Sol, while Free and Go accounts get GPT-6 Luna, which OpenAI describes as tuned for everyday conversation. Intelligent UI works from Instant through Extra High reasoning — but the Pro reasoning option still uses GPT-6 Astra and does not support the new interface system, and the models behind Codex and ChatGPT Work did not change.
Availability is phased: Enterprise access can depend on workspace administrator settings, and rollout can vary by account, region, and app version. OpenAI's help page notes that older ChatGPT desktop apps for Windows and macOS don't support Intelligent UI — the web version is recommended, and phone apps need the latest version. There is no separate usage quota for the interactive features; your plan's normal limits apply.
One honest caveat, from OpenAI itself: the company acknowledges that the model's design judgment still needs work. If the generated interface looks clumsy on a given prompt, that's a known limitation, not a bug in your usage. Try it on a concrete task — a bill splitter, a trip plan, a side-by-side comparison — and judge the tool, not the promise.