ChatGPT Answers Turned Into Charts and Widgets? How to Get Plain Text Back After the GPT-6 Intelligent UI Update (2026)

You asked a simple question about splitting rent. ChatGPT handed you a little calculator with sliders.

Nice, maybe. Not so nice if you needed three lines of text to paste into a group chat, or a table for a spreadsheet, or something a screen reader handles well. Plenty of people want the old wall of text back. It’s not quite available.

Here’s what changed with GPT-6, which switch actually exists, and the workarounds that get you clean, copyable text.

Contents

• What is Intelligent UI, and why did my ChatGPT answers change?
• Can you turn off Intelligent UI in ChatGPT?
• How do I get plain text answers I can copy?
• Why does ChatGPT start answering before it finishes?
• Not Seeing Intelligent UI at All? Six Usual Reasons
• Can you trust the numbers inside a ChatGPT widget?
• Tips
• Warnings
• FAQ
• Sources

What is Intelligent UI, and why did my ChatGPT answers change?

Intelligent UI is a GPT-6 feature in ChatGPT that lets a single answer mix text with visuals and interactive elements such as comparisons, diagrams, calculators, bill splitters, and games, and OpenAI’s ChatGPT release notes dated October 7, 2026 say the model chooses that format automatically.

The rollout began on October 7 for Plus, Pro, Business, and Enterprise in the Chat tab. Free and Go followed from October 8. Enterprise workspaces get it only if the admin settings allow.

When I checked the release notes, I noticed that OpenAI keeps one escape hatch in its own wording: ChatGPT can “still use a simple text answer when that works best.” So text didn’t disappear. The model just decides more often that it shouldn’t be text.

Compiled from OpenAI’s ChatGPT release notes of October 7, 2026. Intelligent UI works from Instant through Extra High reasoning. The models behind Work and Codex did not change.

Can you turn off Intelligent UI in ChatGPT?

You cannot fully turn off Intelligent UI in ChatGPT; as of October 2026, OpenAI’s help article describes a Layout and visuals setting on the web that reduces visual elements but says some may still appear and that it does not switch answers to a text-only mode.

It’s still worth flipping. It shifts the default for every new chat.

Method 1: Turn down visuals on the web

1. Open chatgpt.com in a browser and sign in.
2. Click your profile icon, then Settings.
3. Find Layout and visuals and switch it off.
4. Start a new chat. Old conversations keep whatever format they already rendered.

The setting is documented for the web. If you use the phone app most of the day, change it on the web anyway, then check whether your phone answers look calmer. Don’t count on it carrying over.

Method 2: Use the Pro reasoning option (Pro plan only)

1. Open the model and reasoning menu at the top of a chat.
2. Pick the Pro reasoning option. The release notes say it still runs on GPT-6 Astra and doesn’t support Intelligent UI.
3. Expect slower replies. Astra thinks longer, so it’s overkill for quick questions.

How do I get plain text answers I can copy?

The most reliable way to get plain-text ChatGPT answers after the October 2026 GPT-6 update is to say so in your first message, because OpenAI built Intelligent UI to choose the format per answer and an explicit format request steers that choice.

A single sentence at the top of a chat usually holds for the whole conversation. Something like:

Answer in plain text for this whole chat. No charts, buttons, widgets, or interactive tools. Short paragraphs, and a simple table only if I ask.

Honestly, that’s enough for most people.

If you paste answers into email, Slack, or a document, add the destination. “Format this so I can paste it into Gmail” gets cleaner spacing than “plain text” alone. For spreadsheets, ask for comma-separated values in a code block. A pasted CSV lands in Google Sheets or Excel as real cells; a pasted widget doesn’t paste as anything useful.

Already got a widget? Reply with “Give me that as plain text with the numbers listed.” It costs one more message. Intelligent UI has no separate usage quota, so it comes out of your normal plan limits like any other reply.

Why does ChatGPT start answering before it finishes?

ChatGPT starts answering early because, per OpenAI’s October 7, 2026 release notes, GPT-6 can begin a reply while it is still thinking or using tools and then add more findings to the same answer without another prompt.

This trips a lot of people up. You see a confident first paragraph, copy it, and miss the correction or extra detail that lands a few seconds later.

The fix is boring. Wait until the stop button disappears and the reply stops growing before you copy anything. If the answer includes a chart or tool, scroll back to the top once it settles; added findings can change what came first.

Most people skip this step and regret it.

Not Seeing Intelligent UI at All? Six Usual Reasons

Some readers have the opposite problem. They want the calculators and never see them. Run down this list.

• You’re on the Pro reasoning option. GPT-6 Astra doesn’t do Intelligent UI. Switch to Instant through Extra High.
• You’re using an older desktop app. OpenAI’s help article says the older ChatGPT desktop apps for macOS and Windows don’t support it. Use chatgpt.com in a browser.
• Your phone app is behind. Update ChatGPT from the App Store or Google Play.
• Your workspace admin hasn’t enabled it. Enterprise availability depends on admin settings.
• You turned off Layout and visuals. That’s the whole point of the switch.
• The model chose text. There’s no magic prompt. Ask for the tool by name: “build me a bill splitter for 4 people with a tip slider.”

If the model menu itself looks wrong, our walkthrough of GPT-6.1 Sol not showing in the model picker covers app versions and plan checks in more detail.

Can you trust the numbers inside a ChatGPT widget?

Numbers inside a ChatGPT Intelligent UI widget come from the same GPT-6 model as the text, so a loan calculator or savings chart is only as accurate as the inputs and formula the model chose, and OpenAI’s launch material does not describe a separate verification step.

A slider makes a number feel more official. It isn’t.

I found that the quickest check is to ask, right under the widget, “What formula and inputs did you use?” If the formula is wrong or an input is a guess, you’ll see it in one reply. For money decisions, rerun the result in a spreadsheet or a dedicated calculator before acting on it.

Also check whether sources are still attached. A chart with no Sources button gives you nothing to verify against.

Tips

• Put your format preference in Custom Instructions if you want it in every new chat, not just one.
• Ask for a “markdown table” when you want a grid that pastes cleanly into docs.
• Use the widget to explore, then ask for a text summary to save. Tools inside a chat aren’t easy to reuse elsewhere.
• Bookmark the Layout and visuals setting page on the web so you can flip it back when you actually want a chart.

Warnings

• Don’t copy a reply while it’s still growing. GPT-6 can add findings after the first paragraph.
• Don’t treat a calculator’s output as financial advice. Check the formula first.
• Don’t download “ChatGPT Classic” or “text-only ChatGPT” apps from third-party sites promising to remove the new layout. Unofficial clients ask for your login.
• OpenAI’s help article and release notes don’t fully agree on the Free and Go model name, so expect wording in the app to shift as OpenAI updates its docs.

FAQ

Does Intelligent UI use up more of my ChatGPT message limit?

No. OpenAI says Intelligent UI has no separate usage quota. An answer with a chart or tool counts like any other message on your plan, though asking it to redo a widget as text does use one more message.

Does the Layout and visuals setting work in the ChatGPT phone app?

OpenAI documents the Layout and visuals switch for ChatGPT on the web. Change it there, then watch whether answers in the app change; if not, add a plain text request at the start of each chat on your phone.

Is Intelligent UI the same as ChatGPT Canvas?

No. Intelligent UI builds visuals and small tools inside the answer itself and arrived with GPT-6 on October 7, 2026. Canvas was a separate side-panel editor for documents and code.

Sources

• OpenAI Help Center — Intelligent UI in ChatGPT
• OpenAI Help Center — ChatGPT release notes (October 7, 2026 entry)
• Releasebot — ChatGPT updates mirror, October 2026

Fact-checked against OpenAI’s ChatGPT release notes and help center as of October 12, 2026. Rollout timing and setting names can differ by plan, app version, and region.