OpenAI’s attempt to move ChatGPT beyond the text box reached every consumer tier on 8 October, when GPT-6 and a feature called Intelligent UI rolled out to Free and Go users, a day after paying subscribers. ChatGPT can now answer with charts, forms, tappable buttons and small working tools, and it decides for itself when a reply should be prose and when it should be something to click.

The change matters to B2B marketing teams because ChatGPT is now a research tool buyers use before they reach a vendor’s site. OpenAI says more than 1.2 billion people use the product each week. When a prospect asks it to compare tools in a category or estimate the cost of switching, the reply may now arrive as a rendered comparison or a calculator assembled inside the conversation, built from the information the model found. How a product is represented in that answer is no longer only a question of what the text says.

What does Intelligent UI add to ChatGPT answers?

Intelligent UI lets ChatGPT combine text with graphics, charts, forms, buttons and other interactive elements, choosing the format to suit the question. OpenAI’s announcement, published on 7 October, says the model will still reply in plain text when that is more useful. It can also build small tools on request: OpenAI’s examples include a retirement savings calculator, a bill splitter and a retro game.

The demonstrations lean towards learning and everyday planning. A seven-speed bicycle is broken down into clickable systems, the Monty Hall problem is explained interactively, a road trip comes with map stops and a Sunday roast with a cooking timeline. OpenAI says the interface is drawn from a library of native components and rendered progressively as the model writes, so a chart can appear before the answer is finished, on web and mobile alike.

TechCrunch, which attended a press briefing, reported that users can reduce the number of visuals in their responses in the same way they adjust ChatGPT’s personality settings. OpenAI product manager Aarush Selvan told reporters that ChatGPT had until now been a predominantly text-based interface.

The company is open about the limits. Its announcement says there is still work ahead on the model’s design judgement and on what it can create. None of the published examples has been tested independently, and OpenAI has released no figures on how often ChatGPT chooses an interactive answer over text.

Which ChatGPT plans get GPT-6, and which model powers each?

GPT-6 reaches ChatGPT users through two models. GPT-6 Sol powers the Plus, Pro, Business and Enterprise plans, where rollout began on 7 October; GPT-6 Luna powers the Free and Go tiers, which followed on 8 October. OpenAI describes both as tuned for everyday conversation but has not published a side-by-side account of what each can do. The release completes a staged launch, after OpenAI introduced its first GPT-6 models to paying customers in September.

Coverage is not universal. ChatGPT’s release notes say Intelligent UI works from Instant through Extra High reasoning, but not in the Pro reasoning option, which runs on a separate model, GPT-6 Astra. The update applies only to the Chat tab, and the models behind ChatGPT Work and Codex are unchanged. In Enterprise workspaces access depends on admin settings, so a marketer inside a large company may not see the new behaviour on the same day as a buyer on a free account.

OpenAI’s performance figures are its own and compare GPT-6 only with its predecessor, GPT-5.6. The company says GPT-6 Instant starts answering web-search questions 44% sooner on average than GPT-5.6 Instant. It also says that, in an internal evaluation of everyday agentic tasks, GPT-6 Extra High begins responding as quickly as GPT-5.6 Medium while scoring better than GPT-5.6 Extra High, without naming the benchmark, sample size or score. OpenAI adds that GPT-6 is better at deciding when to search and finds supporting sources more reliably, which bears directly on any company whose product information lives on the open web.

What does the change mean for how B2B products appear in ChatGPT?

Intelligent UI changes how ChatGPT presents an answer, not how it chooses its sources. OpenAI has announced nothing about how the model selects or weights the information behind a reply, and nothing in the release amounts to a ranking change. What shifts is what the buyer sees: a pricing question may come back as a table, a migration question as an estimator, a category question as a chart of options.

Anthropic’s Claude has produced interactive artifacts since 2024, so rendered output inside an assistant is not new. ChatGPT’s reach makes its version the one most buyers are likely to meet, and on the free tier by default.

For marketing teams that already check how ChatGPT describes their category, the release adds a layer to that work. An interactive comparison built from out-of-date pricing or a misread feature page can carry more weight than the same mistake in a paragraph, because it looks like a finished tool. The practical response is familiar: run real buyer questions through both free and paid tiers, record what comes back in each format, and keep pricing pages, documentation and comparison content current and unambiguous.

The same capability is available on the production side. Teams that use ChatGPT for content can ask it for a working calculator or an interactive explainer rather than a text draft, provided someone checks the output for accuracy, accessibility and mobile behaviour before a customer sees it. OpenAI’s own caveat on design judgement applies there too.

OpenAI has not said when Intelligent UI will reach the Pro reasoning option or the Work and Codex models.