On October 7, 2026, OpenAI began rolling out a feature called Intelligent UI that changes the shape of ChatGPT’s answers. Instead of returning only text, GPT-6 can arrange text, graphics, buttons, tables and forms into interactive responses inside the conversation itself. According to TechRepublic, the rollout started with the Plus, Pro, Business and Enterprise plans, with expansion to the Free and Go plans scheduled to begin on October 8.
The underlying idea is simple to explain. Until now, anyone who wanted to compare several scenarios in ChatGPT, for example different household budgets, had to type a new question every time a number changed. With Intelligent UI, users can adjust contributions to a savings plan, explore the parts of a diagram or change the stops of a trip on a map without asking for a whole new answer.
OpenAI says no special prompt is needed to trigger these responses, although users can explicitly ask ChatGPT to create a tool for a particular task. The company cited several examples: a calculator to split a dinner bill, a savings calculator with adjustable amounts, an interactive diagram with selectable parts to understand a concept, and a map with stops and possible detours to plan a road trip.
On the technical side, OpenAI explained that it built Intelligent UI with native, streamable components and a compiler that displays the interface as GPT-6 generates it. ChatGPT can also begin answering while it continues to reason and add information across successive responses, so users see results before the model has finished thinking.
The company paired the launch with a speed figure published on its developer community: GPT-6 Instant began answering web search questions, on average, 44% sooner than GPT-5.6 Instant. TechRepublic stressed an important nuance: the figure measures the time until a response begins, not the time needed to complete it, and it does not say whether an interactive chart is ready any sooner.
OpenAI also said it had improved the model’s ability to decide when to search the web and to find information that supports its answers. These are, in any case, results reported by the company itself rather than independent measurements. Until third parties publish comparable tests, those improvements are best read as vendor claims.
Access conditions come in several layers. According to OpenAI’s Help Center, Intelligent UI works on the web and in supported updated apps when GPT-6 runs at reasoning levels from Instant through Extra High. It is not available at the Pro effort level or in older versions of the macOS and Windows desktop apps. On Enterprise accounts, access can depend on administrators’ settings.
The feature has no separate usage quota, although existing model, tool and plan limits still apply. The rollout also applies to ChatGPT’s chat experience, not to the models powering Work or Codex. Those who prefer less visual answers can turn off the “Layout and visuals” option in the web settings, although OpenAI warns that doing so does not guarantee text-only answers.
The launch fits a broader trend: AI assistants are moving from text boxes to workspaces. If a user can move a slider and watch a result change, the conversation looks more like a spreadsheet or a small custom app than an exchange of messages. For OpenAI, which competes with Google and Anthropic for daily usage time, it is a way to keep users inside ChatGPT.
For people who design digital products, the change also has its own reading. If a general assistant can build a calculator, a map or a form on the fly, tailored to each user’s question, some of the small utilities that exist today as standalone apps or web pages could end up being handled inside the chat. The trade-off is that these generated interfaces do not go through the design, testing and review process that comes with a traditional product, which makes careful verification of their results all the more important.
For corporate IT teams, TechRepublic suggested a practical way to test the feature: comparing hypothetical software subscription costs. A calculator can show how monthly spending changes as the number of users grows, offering a starting point before reviewing actual vendor quotes.
That same usefulness carries a risk. An interactive answer conveys a sense of precision that is not always warranted: the figures, calculations and assumptions behind a model-generated calculator can contain errors. TechRepublic recommended verifying the numbers, especially when comparing costs or making business decisions, and following company policies before entering sensitive information.
OpenAI acknowledges that it is still refining when GPT-6 should generate interactive tools and that it plans to support more types of interfaces. For now, not every question will produce a chart or a form, even when one might be useful. That unpredictability is common in the first weeks of a new feature and will probably shrink as the company gathers usage data.
From a wider perspective, Intelligent UI shows where competition between assistants is heading. Improvements are no longer measured only in reasoning benchmark scores, but also in time to first response, in the ability to present information in a form users can manipulate, and in integration with everyday tasks such as planning a trip or splitting a bill.
In short, OpenAI has turned part of ChatGPT’s answers into small tools that users can manipulate without typing again. The immediate advantage is being able to compare options quickly within a single conversation; the limit, as coverage of the launch points out, is that those results should be treated as a starting point for further research rather than a final answer.
