The OpenAI logo. Image: OpenAI logo via LobeHub Icons (MIT); trademark of OpenAI
OpenAI has started rolling out GPT-6 to everyone who uses ChatGPT, and with it a feature it calls Intelligent UI: instead of a wall of text, answers can now come back as charts, maps, buttons, forms and small working tools built on the spot.
Plus, Pro, Business and Enterprise users get it from today in the Chat tab, and Free and Go users from tomorrow. OpenAI says it is aimed at the “more than 1.2 billion people who use ChatGPT each week”.
What Intelligent UI does #
OpenAI says it trained GPT-6 to “compose responses using text, visuals and interactive elements”, picking the format to suit the question. A comparison might appear side by side, a road trip as stops on a map, and a Sunday roast plan with the cooking times laid out next to the recipe. Explanations can become interactive diagrams where you change an input and see what happens.
You can also ask it to make a tool there and then, and OpenAI’s examples are a savings calculator, a bill splitter for dinner with friends and a game you play inside the chat. When plain text is the best answer, OpenAI says ChatGPT will still just give you text.
Under the hood, OpenAI built a library of ready-made components and a compiler that turns the model’s output into an interface as it streams, so the page fills in while the answer is still being written. It admits there is “still work ahead to improve the model’s design judgment”.
Faster answers, then more thinking #
The other big change is speed. GPT-6 in ChatGPT can give a first answer and then keep thinking, adding detail in follow-up parts rather than making you wait for the whole reply. OpenAI says the model was trained to take “the user’s waiting time” into account.
By OpenAI’s own measures, GPT-6 Instant starts answering questions that need a web search 44% sooner on average than GPT-5.6 Instant, and is better at deciding when to look something up. In an internal test of everyday agent tasks, it says GPT-6 Extra High starts answering as quickly as GPT-5.6 Medium while scoring better than GPT-5.6 Extra High. These are OpenAI’s figures and haven’t been independently checked.
Which GPT-6 you get #
The model behind the new Chat experience depends on your plan. Paid tiers get GPT-6 Sol, while Free and Go users get the smaller GPT-6 Luna, both tuned for everyday conversation. The models behind ChatGPT’s Work mode and Codex aren’t changing, and in Enterprise accounts the rollout depends on admin settings.
OpenAI first released GPT-6 models to paying customers and developers last month, and launched the upgraded GPT-6.1 Sol at DevDay. On safety, it says GPT-6 builds on work done for GPT-6 Astra, resisting jailbreak attempts better, particularly ones spread over several turns, and is better at saying when it doesn’t know something. Details are in its system card.
“Software will adapt to people” #
OpenAI frames Intelligent UI as the start of something bigger. “For decades, people have had to learn how to use software, navigating fixed interfaces designed around particular tasks,” it says. “Instead of people adapting to software, software will adapt to people.”
The launch lands on a busy day for both big labs: Anthropic released its much cheaper Claude Haiku 5.5 a few hours earlier.
Why it matters #
This is the biggest change to how ChatGPT looks since it launched, and it reaches free users within a day. If it works, ChatGPT stops being a chat window and starts acting like an app that builds itself around each question, which puts it in more direct competition with the apps and websites people would otherwise open.