OpenAI rolls out GPT-6 to all ChatGPT tiers with Interactive UI

OpenAI has started rolling out GPT-6 to every tier of ChatGPT, from paid Plus, Pro, Business and Enterprise accounts down to Free and low-cost Go plans. The release arrives with a capability the company calls Intelligent UI, which OpenAI says lets the chatbot "return responses quickly with fully interactive user interfaces." The rollout to paid tiers has already begun worldwide; Free and Go users get access to the new model the same day.

The stakes here are straightforward. Chatbots have so far delivered answers as walls of text, or as text with occasional images. GPT-6 moves ChatGPT toward generating actual software on the fly — a shift that matters for how hundreds of millions of users consume information, and for the competitive race among AI labs to make assistants indispensable rather than merely conversational.

What does Intelligent UI actually do?

OpenAI trained GPT-6 to compose visual, interactive responses directly inside the chatbot. Depending on what you ask, replies can include graphics, tappable buttons, forms, charts and other interactive elements. The model decides what fits: if plain text makes more sense, it can skip visuals entirely.

The company's own example is dinner planning. Ask ChatGPT for help hosting, and the response could include an interactive table showing how many ingredients to buy based on the guest count you provide. It could also attach checklists and follow-up forms you fill in to get more detailed help.

Beyond household tasks, OpenAI positions the feature as a way to understand complex topics through visual aids — and as a tool builder. Users can create specialized utilities for their own needs, such as a calculator for investment decisions or for splitting a bill. You can even use it to recreate retro games inside ChatGPT, like Snake.

OpenAI framed the ambition in its announcement: "Instead of fitting every question into the same conversational format, ChatGPT can combine UI, data, and actions to create entirely new ways for people to explore, understand, and interact with information."

How does it answer while still thinking?

GPT-6 breaks from the reasoning-model pattern that has dominated recent releases. Other reasoning models decide their output before responding, which produces wait times on complex tasks. OpenAI gave GPT-6 the capability to answer while it is still thinking, explicitly to shorten those delays.

The company also trained the model to compose responses across multiple messages. Each message adds useful information as it arrives, and the final output is designed to be cohesive rather than fragmented.

That combination — streaming reasoning plus multi-message composition — is what makes dynamic interfaces practical. A response can begin building an interactive element while the underlying computation continues.

Which model do you actually get?

Not every tier receives the same GPT-6. Paid subscribers on ChatGPT Plus, Pro, Business and Enterprise get the release powered by GPT-6 Sol, the stronger variant. Free users and those on the low-cost Go plan get GPT-6 Luna, which OpenAI says isn't as capable as Sol.

The split continues OpenAI's tiered-model strategy: same feature set, different capability ceilings depending on what you pay.

Why this release matters

The rollout signals where OpenAI thinks chat interfaces are heading. If responses become interactive applications — tables, forms, charts, playable games — ChatGPT stops being a question-and-answer box and starts competing with general-purpose software tools. The investment calculator and the bill splitter are small examples of that logic: the assistant doesn't describe a tool, it builds one on demand.

The tiering also sets up a measurable capability gap between Sol and Luna that users and developers will probe as access spreads. With Free-tier rollout landing today, expect the first wave of hands-on comparisons between the two variants — and the first recreations of Snake — within hours.