OpenAI opened GPT-6 to all ChatGPT users on October 7, together with a capability it calls Intelligent UI. The first GPT-6 models landed in September, for paying customers only. Now a dedicated GPT-6 goes to the more than 1.2 billion people who use ChatGPT every week.
The answer isn’t only text anymore
Intelligent UI means ChatGPT can assemble a response from text, graphics, tappable buttons, forms and charts, usable right inside the conversation. The model picks the shape per question. A comparison lands side by side, an explanation becomes an interactive diagram, and when plain text is the most useful answer, you get plain text.
The examples in the announcement: the timing for a Sunday roast sits next to the recipe, the stops on a road trip appear on a map with notes on what’s worth a detour, and the build of a 7-speed bike gets pulled apart into tabs filtered by assembly.
A compiler renders while the model writes
Under the hood there are two pieces: a library of native, streamable components, and a compiler that processes the interface as the model generates it. So the interface appears progressively instead of waiting for the full response. The library gives every answer the same design foundation; the model decides how the pieces fit.
Training came through evaluating the interfaces the model produces for clarity, usefulness and completeness. OpenAI says outright that the model’s design judgment still needs work.
Answers start before the thinking is done
The second change is less visible and probably matters more day to day: GPT-6 can begin answering while it keeps thinking. With reasoning models you used to wait for the thinking to finish before anything appeared.
Now ChatGPT interleaves the two. The model factors in your waiting time and builds the answer across several partial responses, each adding something. OpenAI’s numbers: GPT-6 Extra High starts answering as fast as GPT-5.6 Medium while scoring better overall than GPT-5.6 Extra High. On questions that need web search, GPT-6 Instant starts answering 44% sooner on average than GPT-5.6 Instant.
Who gets what
The rollout started globally on October 7 for Plus, Pro, Business and Enterprise in the Chat tab. Free and Go follow a day later. On Enterprise, the admin setting decides.
Two models sit underneath: GPT-6 Sol for the paid tiers, GPT-6 Luna for Free and Go. Both are tuned for everyday conversation. Work and Codex stay on their current models — this update only touches Chat.
The interface is the real argument
On safety OpenAI sticks to the familiar phrasing: better resistance to multi-turn jailbreaks, retrained safeguards against cyber, bio and violence misuse, fewer unnecessary refusals. Without the system card in hand, little of that is checkable.
What’s more interesting is what Intelligent UI quietly shifts. An answer delivered as an interface is harder to quote, copy and contradict than a paragraph of prose. You can operate it, but you can’t read it line by line. OpenAI sells this as the end of people adapting to software. It’s also the end of an answer just sitting there, where you could check it sentence by sentence.