GPT-6 Reaches Every ChatGPT Tier This Week. It Can Answer You With a Bill Splitter.

Ask ChatGPT to split a dinner bill four ways and you get a paragraph of arithmetic back. Starting this week it can hand you a bill splitter instead, a small tool with fields you type into, sitting inside the conversation.

That capability is called Intelligent UI, and OpenAI announced it on Wednesday alongside the rollout of GPT-6 to every ChatGPT tier. GPT-6 Sol starts reaching Plus, Pro, Business and Enterprise accounts on October 7 in the Chat tab, with GPT-6 Luna arriving for Free and Go users on the 8th. Paid customers got the first GPT-6 models last month. This is everyone else's turn.

ChatGPT picks the format now

The pitch, in OpenAI's words, is "Intelligent UI, a new capability that allows ChatGPT to answer quickly with fully interactive user interfaces." In practice, answers can come back as visuals, tappable buttons, forms, charts and small interactive tools rendered inline rather than as prose. The model decides which shape fits the question. Ask it to compare two things and you may get them side by side. Ask it to explain something and you may get a diagram you can poke at. Ask it something simple and you still get a sentence.

The examples OpenAI gave are meal-timing plans, road-trip maps, interactive statistics explainers, and small utilities like a bill splitter or a savings calculator. A schedule beats a paragraph about a schedule.

How it draws that fast

Two pieces make it work. OpenAI built a library of native, streamable components, plus a compiler that renders the interface progressively as the model generates it, so you are not watching a spinner while a chart assembles itself somewhere. ChatGPT can also begin answering while it is still reasoning, instead of waiting for the reasoning to finish.

That second one changes the feel of using it more than the components do. Watching a model think before it says a word is the most tiring part of talking to one.

The 44% number is OpenAI's own

OpenAI says web-search answers now start about 44% sooner on average than they did with GPT-5.6 Instant. OpenAI also did the measuring. Nobody outside the company has timed it.

The safety claims come with the same asterisk. OpenAI reports that GPT-6 holds up better against attempts to bypass its safety training and has a sharper sense of its own limits, and it credits some of that to work done on GPT-6 Astra. Those are OpenAI's own evaluations too.

Who gets it, and who does not

Sol goes to Plus, Pro, Business and Enterprise in the Chat tab starting October 7. Enterprise users may not see it right away, since access there depends on what their workplace admin has turned on. Luna goes to Free and Go on October 8. The Work and Codex experiences are not changing in this release, and OpenAI announced no pricing changes.

OpenAI says GPT-6 is built for more than 1.2 billion weekly ChatGPT users, which is the scale that justifies building a component library and a compiler to go with it.

"Instead of people adapting to software, software will adapt to people," OpenAI writes. A chatbot that answers with a working calculator when a calculator is what you wanted is genuinely new. Over the next few weeks the question is how often it picks wrong, and whether a one-line answer starts arriving wrapped in a widget you have to tap through.