ChatGPT can now answer with a working interface instead of a wall of text. On Wednesday, OpenAI announced Intelligent UI, a feature that mixes text, visuals and interactive elements right inside a reply, and began rolling out new GPT-6 models built for everyday conversations.
What Intelligent UI does
Instead of one long paragraph, ChatGPT can return tappable buttons, forms, charts and diagrams. The format follows the question. A comparison might show up side by side, an explanation might become an interactive diagram, and a simple question still gets plain text.
OpenAI’s examples include more visuals, a recipe widget that adjusts ingredient amounts for the number of guests, checklists, built-in maps and multiple-choice follow-ups that narrow down results. ChatGPT can also build small tools on the spot, such as a calculator, a bill splitter or a simple game.
Under the hood, OpenAI says the feature relies on a library of native components that can be streamed, with a compiler processing the interface while the model is still generating it.
The GPT-6 models behind it
OpenAI introduced the first GPT-6 models last month. The new ones are tuned for regular ChatGPT chats:
- GPT-6 Sol goes to Plus, Pro, Business and Enterprise subscribers, starting October 7, 2026.
- GPT-6 Luna reaches Free and Go users the next day, October 8, 2026.
GPT-6 can also start replying while it is still reasoning and add detail as it goes, which OpenAI says makes answers feel quicker.
The change applies to the Chat tab only. The models used for Work and Codex are not changing.
What we don’t know yet
The coverage doesn’t spell out usage limits by plan, and OpenAI calls the rollout global and gradual, so your account may get it a little after someone else’s. No new subscription prices came with the announcement.
Why it matters
Until now, asking ChatGPT to compare two options meant reading a table in text. With Intelligent UI, the answer is meant to be the tool itself. That points toward assistants that do things rather than just explain them, and it puts pressure on rivals like Gemini. If you want background on why these systems sometimes get things wrong, see our explainer on what a language model is. You can also read about the ads coming to the image generator.
