ChatGPT Intelligent UI is the reason your answers suddenly contain charts, maps, buttons and small working tools. OpenAI switched it on with GPT-6 in the Chat tab on 7 October 2026, for Plus, Pro, Business and Enterprise first and for Free and Go from 8 October. Ask how to split a dinner bill and you may get a bill splitter you can type into, not a paragraph of arithmetic.

Some people love it. Others want their plain text back. The catch is that OpenAI's off switch only turns the visuals down, not off.

This guide covers what Intelligent UI builds and which plans get it. It shows how to trigger a tool on purpose, how to get text-only answers and why the feature may not show up. It ends with how Intelligent UI compares with Gemini's visual layout and Claude Artifacts.

The Key Takeaways

  • What it is: a GPT-6 feature that lets ChatGPT answer with charts, maps, forms, tappable buttons and mini tools like calculators, bill splitters and games.
  • Who gets it: every plan in the Chat tab, from Instant through Extra High. The Pro thinking level runs on GPT-6 Astra and does not get it.
  • The off switch is partial: Settings, then Layout and visuals, on the web. OpenAI says some visuals may still appear and there is no text-only mode.
  • The old Mac app is left out: OpenAI says the older ChatGPT desktop apps for macOS and Windows do not support it and points you to the web.
  • No extra cost: Intelligent UI has no separate usage quota. Your normal plan limits apply.

What Is ChatGPT Intelligent UI?

Del editor

Todos los modelos de IA en una sola app

Fello AI reúne GPT-6, Claude 5.5, Gemini 3.8, Grok 4.7 y más en una sola app nativa para Mac y iPhone.

¡Descárgala ahora!

ChatGPT Intelligent UI is a GPT-6 capability that lets one answer combine text, visuals and interactive elements, with ChatGPT choosing the format for each question. OpenAI's announcement lists graphics, tappable buttons, forms, charts and interactive experiences you use directly in the conversation. The ChatGPT release notes for 7 October put it plainly: "It can choose the format automatically, and still use a simple text answer when that works best."

What Intelligent UI Builds

Intelligent UI builds three kinds of answer, according to OpenAI's launch post. The first is a more visual everyday answer: a Sunday roast recipe with the cooking timeline beside it, or the stops on a road trip placed on a map. The second is an interactive explanation, where you change an input and watch the result move. The third is a small tool built on request. OpenAI's examples are a calculator for how savings could grow, a bill splitter for dinner with friends and a game you play inside the chat.

OpenAI's help article on GPT-6 adds that "no special prompt is required". ChatGPT decides on its own when a chart or a tool beats a paragraph. You can still ask for one directly, which is covered in the prompt section of this guide.

How Intelligent UI Works

Intelligent UI runs on a library of native, streamable components plus a compiler that assembles the interface while GPT-6 is still writing. That is why a chart or a form appears piece by piece instead of popping in at the end. OpenAI says it trained GPT-6 to judge layout, visuals and when interactivity helps, and it scored the interfaces the model built for clarity, usefulness and completeness.

Intelligent UI is also not ChatGPT's first visual feature. The release notes show interactive learning modules for 70+ math and science topics on 10 March 2026 and interactive bar, line, pie and scatter charts on 8 June 2026. What changed on 7 October is scope: GPT-6 can now build the interface for almost any question, not only for a fixed list of topics.

Sam Altman's reaction on launch day shows how hard OpenAI is leaning on the change:

Which ChatGPT Plans Get Intelligent UI?

Every ChatGPT plan gets Intelligent UI in the Chat tab, but only at the Instant, Medium, High and Extra High thinking levels. The Pro level is the exception. OpenAI's release notes say "The Pro reasoning option continues to use GPT-6 Astra and doesn’t support Intelligent UI." If you want to know what Astra itself is, our GPT-6 Astra guide covers its benchmarks and price.

PlanModel behind ChatThinking levels with Intelligent UIRollout
Free and GoGPT-6 Luna (release notes) or GPT-5.6 Luna (help article)InstantFrom 8 October 2026
PlusGPT-6 SolInstant, Medium, High (Plus has no Extra High)From 7 October 2026
ProGPT-6 Sol; Pro level uses GPT-6 AstraInstant to Extra High, not ProFrom 7 October 2026
BusinessGPT-6 SolInstant to Extra High, not ProFrom 7 October 2026
EnterpriseGPT-6 SolInstant to Extra High, not ProDepends on workspace admin settings

The Free and Go row carries a real contradiction. The 7 October release notes and OpenAI's launch post both say Free and Go use GPT-6 Luna. OpenAI's help article on GPT-6, updated the same day, still says Free and Go users receive GPT-5.6 Luna. Until the two documents agree, treat the Free-tier model as unconfirmed. Our guide to what ChatGPT Free includes tracks the rest of the Free plan.

Intelligent UI has no separate usage quota. The help article says "Intelligent UI has no separate usage quota." A chart or a calculator costs the same as any other message on your plan.

How to Turn Off ChatGPT Intelligent UI

You cannot fully turn off ChatGPT Intelligent UI. You can only turn it down. OpenAI documents one setting, on ChatGPT on the web:

  1. Open chatgpt.com in a browser and sign in.
  2. Open Settings.
  3. Find Layout and visuals and switch it off.
  4. Start a new chat and ask your question again.

The help article is explicit about the limit: "Some visual elements may still appear; this setting does not switch the entire answer experience to a classic or text-only mode." OpenAI documents the switch for ChatGPT on the web only and names no equivalent in the iPhone or Android app.

Get Plain Text Back in One Chat

For a plain-text answer in one conversation, ask for it in the prompt. ChatGPT is built to fall back to text when text works best, so an explicit request tips that decision your way. Paste this at the start of a chat:

Answer in plain text only for this whole conversation. No charts, maps, buttons, forms, widgets or interactive tools. Use short paragraphs and, where needed, a simple bulleted list or a Markdown table.

Two more ways get you text without touching settings. Pick the Pro thinking level if your plan has it, since GPT-6 Astra does not use Intelligent UI. Or select GPT-5.6 Sol explicitly in the model menu: OpenAI documents Intelligent UI only for GPT-6, and the help article says explicit GPT-5.6 Sol selections stay on GPT-5.6.

How to Get ChatGPT to Build a Tool

To get a tool, ask for the tool by name and give it the inputs it needs. Intelligent UI picks a format on its own, but a request that names the thing ("a calculator", "a checklist", "a map") removes the guesswork. OpenAI's ChatGPT account pitched the feature on launch day as a way to create tools on the spot:

These prompts mirror the examples OpenAI shows. Swap the bracketed parts for your own numbers.

Build me a bill splitter for dinner with [5] friends. Total is [$240] including tax. Let me set a tip percentage and mark who skipped drinks.
Make a savings calculator. Starting balance [$2,000], monthly deposit [$150], and let me change the interest rate and the number of years with sliders.
Plan a [4]-day road trip from [Denver] to [Moab] and show the stops on a map with a note on which ones are worth a detour.
Explain how [compound interest] works with an interactive diagram where I can change one input and see what happens.

If the answer comes back as plain text, reply with show that as an interactive tool. Check that you are on a thinking level from Instant to Extra High, not Pro.

Why ChatGPT Intelligent UI Is Not Showing

ChatGPT Intelligent UI usually fails to show for one of six reasons, and most of them are settings, not bugs.

You are on the Pro thinking level. Pro uses GPT-6 Astra, which does not support Intelligent UI.

You use the older desktop app. OpenAI's help article says "The older ChatGPT desktop apps for macOS and Windows do not support these capabilities; use ChatGPT on the web instead." OpenAI rebuilt its desktop app on 9 July 2026 and said the previous one may stay installed as ChatGPT Classic. Our ChatGPT Mac app guide explains which app you have.

Your phone app is out of date. OpenAI asks you to update the mobile app to the latest version.

You picked GPT-5.6 Sol. An explicit GPT-5.6 Sol selection stays on GPT-5.6, and Intelligent UI is documented for GPT-6.

The rollout has not reached you. Free and Go started on 8 October 2026, and Enterprise depends on your admin.

You turned Layout and visuals off. The setting reduces visual elements across your answers.

Some inconsistency is normal too. In OpenAI's developer forum, users who saw automatic interactive answers before the launch, in late September, reported them appearing and vanishing. One wrote: "It’s sometimes there and sometimes not." ChatGPT also decides per answer whether a visual helps, so a plain answer is not always a sign that something is broken.

Where Intelligent UI Falls Short

Intelligent UI's biggest weakness is control: ChatGPT decides the format, and the only documented setting turns visuals down rather than off.

OpenAI says as much about GPT-6's taste, writing that "There's still work ahead to improve the model's design judgment" and expand what it can create. Expect the occasional chart nobody asked for.

Sources are the second open question. Search Engine Journal noted that OpenAI's announcement does not say how sources appear when an answer becomes a chart, a comparison or a tool. If you depend on citations, check that the Sources button is still there before trusting a number inside a widget.

The third gap is what happens to a tool after the chat. None of OpenAI's launch documents describes saving, sharing or exporting a tool ChatGPT builds. As documented, a bill splitter lives in the conversation that made it. That is the clearest line between Intelligent UI and Claude Artifacts, and it is a different feature from ChatGPT Canvas, which OpenAI removed on 28 May 2026. Our ChatGPT Canvas explainer covers what replaced it.

ChatGPT Intelligent UI vs Gemini and Claude Artifacts

ChatGPT Intelligent UI is the broadest of the three: it is on by default in ordinary answers on every plan. Google's Gemini app got there first with opt-in experiments, and Anthropic's Claude Artifacts build things you keep. The table compares what each vendor documents.

FeatureChatGPT Intelligent UIGemini visual layout and dynamic viewClaude Artifacts
Launched7 October 2026, with GPT-618 November 2025, as Gemini Labs experimentsAvailable now; Anthropic's help page gives no launch date
How it startsAutomatic in any answer, or on requestChosen from the tool menu, for subsets of usersOn request, or when Claude judges the content fits
Where it appearsInside the chat answerInside the responseIn its own window beside the chat
Keep and share itNot described by OpenAINot described by GoogleSaved in the Artifacts tab, shareable
Who gets itAll plans, Instant to Extra HighRolling out to subsets of usersAll plans can create artifacts in a chat
Turn it downLayout and visuals, web only, not text-onlyLeave the Labs options unselected; the May 2026 richer answers are automaticAsk for text, or do not ask for an artifact

Gemini Visual Layout and Dynamic View

Gemini's visual layout builds a magazine-style answer with photos and interactive modules, and dynamic view codes a custom interface for one prompt. Google's release notes introduced both on 18 November 2025 as experimental Labs features in the tool menu, rolled out to different subsets of users. On 19 May 2026 Google added richer answers to the main Gemini chat for specific topics: zoomable multi-layer images and short narrated video overviews, in English, on the Pro model. Our Gemini Canvas guide covers Google's separate editor for longer work.

Claude Artifacts

Claude Artifacts are the opposite design: a thing you keep, not a format that passes. Anthropic's help article says an artifact "opens beside your conversation", and everything Claude makes lands in the Artifacts tab. You can edit it, share it and come back to it later, which Intelligent UI does not document. The trade-off is effort. Artifacts suit a dashboard or a small app you will reuse. Intelligent UI suits a tool you need for thirty seconds.

If You Want the Answer, Not the Interface

Fello AI lets you pick the model instead of letting one vendor pick the layout. It puts GPT, Claude, Gemini, Grok, DeepSeek and more in one native app for Mac, iPhone and iPad. You can ask the same question of several models side by side and keep whichever answer reads best. And on a Mac it runs as a native app, not a browser tab.

The Bottom Line

Intelligent UI is a real upgrade for questions that are about numbers, places or steps. A bill splitter beats a paragraph about a bill. But it is OpenAI's choice, not yours, and the Layout and visuals switch only turns it down. If you want text, ask for text in the prompt or use the Pro level. If you want a tool, name it. If you want something you can keep and share, that is still Claude Artifacts' job, not ChatGPT's.

FAQ

Is ChatGPT Intelligent UI free?

Yes. OpenAI began rolling Intelligent UI out to Free and Go users on 8 October 2026. It has no separate usage quota, so it uses your normal plan limits.

Can I turn off Intelligent UI completely?

No. Switching off Layout and visuals in Settings on the web reduces visual elements, but OpenAI says some may still appear and there is no text-only mode. Ask for plain text in the prompt to get closer.

Does Intelligent UI work in the ChatGPT Mac app?

Not in the older ChatGPT desktop apps. OpenAI says the older apps for macOS and Windows do not support it and tells you to use ChatGPT on the web instead.

Why does ChatGPT Pro not show Intelligent UI?

The Pro thinking level runs on GPT-6 Astra, which does not support Intelligent UI. Switch to Instant, Medium, High or Extra High to get it.

Is Intelligent UI the same as ChatGPT Canvas?

No. Canvas was a side-panel editor that OpenAI removed on 28 May 2026. Intelligent UI builds visuals and tools inside the answer itself, and it arrived with GPT-6 on 7 October 2026.