OpenAI moves ChatGPT model selector into prompt bar, adds hidden Pro option
Key Points
- OpenAI moves the model selector to the prompt bar for Plus, Pro and Business users.
- Instant (GPT‑5.3) offers fast answers; Thinking (GPT‑5.5) provides deeper reasoning.
- Configure menu lets users pick older models, toggle auto‑selection, and choose Extended thinking time.
- Pro option unlocks the most powerful GPT‑5.5 engine, available only to Pro subscribers.
- Instant uses less data‑center power, offering a greener alternative for simple queries.
- Free and Go plan users see no UI change; mobile app remains unchanged.
OpenAI has quietly revamped the ChatGPT interface for Plus, Pro and Business users, placing the model selector—Instant, Thinking and Configure—directly beside the prompt box. The change makes it easier for subscribers to pick the right version of GPT for their tasks, with Instant delivering speed, Thinking offering deeper reasoning, and a hidden Pro tier unlocking the most powerful GPT‑5.5 engine. A new auto‑select switch and an extended thinking mode also appear in the Configure menu, giving users finer control over performance and resource usage.
OpenAI rolled out a subtle but significant update to the ChatGPT web interface on Monday, moving the model selection menu from an obscure drop‑down at the top of the screen to a button right next to the prompt box. The shift affects all users on the Plus, Pro and Business plans, and it puts three primary choices—Instant, Thinking and Configure—front and center.
Instant, which runs on GPT‑5.3, now appears as the default fast‑response option. Users who need a quick answer—such as the average speed of a jumbo jet—can stay on Instant and get a reply in under ten seconds. Thinking, powered by the newer GPT‑5.5 engine, is designed for more complex queries that benefit from longer reasoning, such as market‑trend analyses. The interface highlights that Thinking can keep track of prior steps better and often prefaces its answer with a brief rationale.
The Configure menu reveals a deeper set of controls. Subscribers can manually select older model versions dating back to GPT‑5.4, GPT‑5.2 and even the older o3 series. An auto‑select toggle—on by default—lets the system decide which model to use based on the question’s difficulty. Below the auto‑select switch, a hidden setting appears when users hover over Thinking: a choice between Standard and Extended thinking time. Selecting Extended pushes the model to spend more processing cycles on the prompt, extracting the maximum of GPT‑5.5’s capabilities.
At the bottom of the Configure list sits the Pro option, the most capable GPT‑5.5 configuration OpenAI currently offers. This tier is reserved for Pro subscribers and is aimed at “hardest tasks and long‑running workflows.” The Pro mode is not visible to Free or Go plan users, who continue to see the old top‑of‑screen menu.
OpenAI also noted an environmental angle. Instant’s lighter processing footprint uses less data‑center power and water for cooling, akin to an eco‑mode on a dishwasher. While the company stopped short of quantifying the savings, the implication is that users who prioritize speed can also reduce their indirect carbon impact.
The redesign mirrors Google’s Gemini interface, which already places its model selector beside the input field. By aligning ChatGPT’s UI with Gemini’s layout, OpenAI appears to be standardizing the user experience across competing AI chat platforms.
For users on the Free or Go plans, the update does not affect the current layout; they still see the legacy menu at the top of the screen. The mobile app version of ChatGPT also retains its previous design, though OpenAI has not confirmed whether these changes will eventually roll out to those platforms.
Industry observers expect the more visible selector to increase usage of the different model tiers, especially Thinking and Pro, which could lead to higher satisfaction for power users and better overall answer quality. OpenAI has not disclosed any pricing adjustments tied to the new UI, but the Pro tier remains exclusive to paying subscribers.