OpenAI

GPT-6 for Everyone: ChatGPT Now Answers With Buttons, Maps and Mini-Apps

co-produced by AI08. October 2026, 14:00

Smartphone showing the ChatGPT start screen next to a laptop — ChatGPT on a smartphone. © Tim Witzdam / Unsplash

Goodbye, wall of text: OpenAI is bringing its new GPT-6 model to ChatGPT for everyone, including free users. Along with it comes a new interface called Intelligent UI, which lets ChatGPT fill its answers with graphics, buttons, forms and small tools. According to OpenAI, the rollout to paying customers began yesterday, and the free version follows today.

Sol for Paying Users, Luna for Everyone Else

Not all GPT-6 is the same. Plus, Pro, Business and Enterprise customers get GPT-6 Sol in chat, while the free version and the low-cost Go plan run the smaller GPT-6 Luna. OpenAI had already released both models in September for ChatGPT Work, Codex and its API, and nothing changes there. In chat, GPT-6 replaces the previous default, while GPT-5.6 Sol remains available separately. The most powerful tier, GPT-6 Astra, still sits behind the Pro thinking mode.

It is also supposed to be faster: According to OpenAI, GPT-6 can start answering before it has finished thinking or searching the web. For questions that involve web search, GPT-6 Instant starts answering 44 percent sooner on average than its predecessor, the company says. An internal evaluation also shows that GPT-6 gets to the heart of difficult questions more often, according to OpenAI, though it did not provide figures.

Clickable Recipes, Calculators on Demand

The real eye-catcher is Intelligent UI. GPT-6 decides for itself what an answer should look like and turns to interactive elements when they help. A few examples:

Recipes: A recipe widget automatically scales ingredient quantities to the number of guests.
Planning: Checklists, embedded maps and multiple-choice follow-up questions.
Tools on demand: Users can have ChatGPT build a calculator, a bill splitter for roommates or a simple game in seconds.

Technically, OpenAI relies on a library of ready-made components that the model assembles as it writes. Where plain text works better, ChatGPT is supposed to stick with text. Users who don’t want the visuals can switch them off on the web under “Layout and visuals,” though OpenAI says some graphics may still appear. Intelligent UI is rolling out gradually on the web and in updated apps, but not in the Pro thinking mode or the older desktop apps for macOS and Windows.

More Show Than Substance?

Not everyone is convinced. Wired headlined its take by calling Intelligent UI more show than tell. Search Engine Journal points to another consequence: If ChatGPT builds calculators and converters right in the chat, websites offering those tools will lose visitors. It is also unclear how sources will be displayed alongside graphics and comparisons.

The launch matters for OpenAI. Recently, Anthropic’s smaller Sonnet 5.5 overtook GPT-6 in rankings, and plans for GPT-6.1 were halted. Google, meanwhile, is going the opposite way with free users and is limiting the free version of Gemini to its smallest model, Flash-Lite, starting tomorrow. Generative interfaces are not unique to OpenAI, either: Google is also building answers with interactive elements into its AI Overviews.