OpenAI launched on Wednesday, October 7, the most ambitious version of ChatGPT since the introduction of Canvas. The new update, internally dubbed "Intelligent UI," transforms the conversation experience into an interactive multimodal experience, filling responses with diagrams, data charts, clickable buttons, and even personalized calculators that users can manipulate directly.
The shift represents a significant departure from the pure text interface that defined ChatGPT since 2022. While competitors — Claude, Gemini, and Perplexity's own interface — focused on speed and reasoning improvements, OpenAI chose to reimagine the interaction format. The result is a chatbot that doesn't just answer questions, but builds live visual tools.
The new interface arrives alongside the GPT-6 Sol and Luna models, the fastest and most efficient versions of the GPT-6 family announced by the company in early September. The launch marks the moment when ChatGPT stops being essentially a text terminal and becomes a visual workspace, with elements that respond to user interactions in real time.
What Intelligent UI does differently
When sending a message to ChatGPT, users will now see elements that go far beyond formatted text. Ask the model to explain how airplane wings generate lift, and it won't just describe the phenomenon — it will generate a series of illustrative diagrams with arrows, labels, and layers of information that can be zoomed or toggled. Ask the model to create a comparative sales data chart, and it won't just display a static PNG, but an editable chart with interactive axes that users can resize, label, and export.
According to Aarush Selvan, product manager at OpenAI, the logic behind the update is that "the most useful responses don't always consist of text alone." In internal demos, the team showed a student asking how to teach a game of Mahjong: ChatGPT generated diagrams of each tile, organized in navigable categories, with buttons to explore rules and strategies visually.
The integration with GPT-6 Sol and Luna also brought performance improvements that were previously impossible with the visual interface. The diagrams and charts now render in real time, without the asynchronous generation delay that characterized OpenAI's earlier attempts at visual interfaces — such as Canvas, launched in 2026, which remained focused on text and code editing.
Why now?
OpenAI's answer to the question "why visual?" involves two factors: capability and competition. GPT-6, with its 1 million token context window and more advanced reasoning capabilities, has the cognitive depth necessary to generate coherent and functionally useful visual elements. But the true catalyst is competitor movement. Anthropic's Claude introduced interactive cards in 2025, Google's Gemini now generates live spreadsheets and dashboards, and Perplexica and Perplexity pushed visual results starting in early 2026.
Sam Altman, CEO of OpenAI, did not comment directly on the Intelligent UI launch day, but in an interview with Wired this week, stated that "pretty soon billions of people a day will be talking to ChatGPT" — implying that the interface needs to scale beyond text to serve a user base that uses the tool not just for questions, but for work, learning, and visual decision-making.
What the new interface is NOT
The Intelligent UI is not an AI that is inherently smarter — it's the same GPT-6 architecture, but with radically different formatted output. Image generation has improved (the ChatGPT Images 2.0, launched in April, already produced more detailed results), but the innovation here is interactivity: buttons that expand information, sliders that adjust parameters, charts that change when you hover over them.
It's also not a direct response to Microsoft's Copilot. Although Microsoft announced a deeper integration of Copilot with Windows 11 and files on the same day, the focus of Intelligent UI is ChatGPT as an individual-use platform, not as an operating system. OpenAI is building the ecosystem, not the OS.
Limitations and risks
The complexity of Intelligent UI brings real challenges. Interactive graphics and dynamic diagrams consume more CPU and GPU resources than pure text, which can impact the experience on older devices. Additionally, personalizing each visual element generates more tokens per interaction, increasing API costs for businesses using the model via the Codex API.
OpenAI has not yet revealed whether Intelligent UI will be available for all plans or if it will be an exclusive Plus/Team feature. With the additional computational cost, there's a risk that the functionality will be restricted to higher-value plans — which would create a divide between text users and visual users, something the company tried to avoid with Canvas.
What to expect in the coming months
If reception is positive, we'll likely see rapid expansions: interactive data tables, event timelines, decision flowcharts, and even mini-applications embedded in conversations. OpenAI has a history of rolling out features gradually — Canvas was rolled to Plus first, then Team, and only later to Enterprise. Intelligent UI will likely follow the same pattern.
The question to watch is: with visual interfaces becoming increasingly sophisticated in ChatGPT, how long until the very notion of "chat" seems archaic, and the tool becomes more like a visual work assistant than a chatbot?
Sources: The Verge, TechCrunch, Wired
✓ Independent sources cross-checked and verified before publishing
