Model: ChatGPT
On 10 June 2026, OpenAI reworked ChatGPT’s model picker into a simpler set of reasoning tiers — Instant, Medium, High and Extra High — retired the old “Thinking” labels, and gave Instant the ability to step up to Medium on its own. It looks like a tidy UX change, but it quietly shifts the level of reasoning behind the answers most people get.
Source: OpenAI ChatGPT release notes — Simplified controls in the model picker (10 June 2026)
What changed in the model picker
ChatGPT’s model picker now presents a single, plain-language scale built around the balance of speed and reasoning effort rather than a list of named “Thinking” modes. The options are Instant, Medium, High and Extra High, with Pro Standard and Pro Extended remaining available on Pro plans.
The picker has also moved to where people actually look for it: at the top of the conversation on iOS and Android, and directly in the message composer on the web. The update is rolling out to Plus and Pro users across web, iOS and Android globally.
The old tiers versus the new tiers
The relabelling is mostly a one-to-one mapping of the previous Thinking modes, with one removal:
Instant stays Instant.
Thinking — Standard becomes Medium.
Thinking — Extended becomes High.
Thinking — Heavy becomes Extra High.
Thinking — Light is removed.
Pro Standard and Pro Extended remain under Pro.
Old “Thinking” modes → new reasoning tiers
Try it: slide the query from simple to complex
A simple question resolves at Instant — the fastest tier answers directly.
On paper this is a renaming exercise. In practice, dropping the lightest reasoning tier and renaming the rest changes the menu people choose from — and the floor for how much reasoning a “quick” answer gets.
Why the reasoning tier shapes the answer
The reasoning tier behind a response is not a cosmetic setting. A fast, low-effort answer and a high-effort one can differ in how thoroughly the model weighs options, whether it reaches for and cites sources, and how many brands or products it surfaces when someone asks for a recommendation. The same question, run at Instant versus High, can return a noticeably different shortlist.
That matters for any brand trying to understand how it shows up in AI answers. When the default level of reasoning shifts — even by a single step — the set of sources and names that make it into the response can shift with it. Simplifying the picker does not change the recommendation logic, but it does change which tier most people land on.
The quiet headline: Instant can now escalate to Medium
The most consequential detail is not the renaming — it is that users can let Instant auto-switch to Medium when a task needs more reasoning, a behaviour you can turn on or off under General > Settings. With Thinking — Light gone, the practical baseline for everyday questions edges upward: a query that once resolved at the lightest tier may now be answered with Medium-level reasoning instead.
A higher reasoning floor tends to mean more considered, more source-grounded answers for ordinary questions — the kind where a brand is either mentioned or quietly left out. That is exactly the layer where AI visibility is won or lost, which is why this update is worth watching rather than waving through.
It also follows the direction of travel we covered when the GPT-5.5 Instant model became ChatGPT’s default — OpenAI continuing to tune what “fast” actually means for the average user.
What we’re watching as the new picker rolls out
The clearest signals will appear in how answers differ across the tiers in the first few weeks. Early markers worth following: whether Medium-tier responses cite more sources than the old light answers did, whether the brands named in recommendation queries change as Instant escalates to Medium, and how consistent answers stay when the same prompt is run at different reasoning levels.
At reconnAI we monitor how the major models answer as these controls change — how decisively they respond, which sources they lean on, and which names surface in the answer. The simplified picker is now part of that tracking.
How reconnAI tracks changes like this
reconnAI tracks how the leading AI models — ChatGPT, Claude, Gemini, Perplexity, Copilot and Google AI Overview — answer questions across regions, and re-baselines that tracking whenever a platform changes how its answers are generated. If you want to understand how ChatGPT represents your brand under the new reasoning tiers, you can get in touch with our team or see how AI visibility tracking works.
About reconnAI
reconnAI tracks how the major AI models represent topics and sources across ChatGPT, Claude, Gemini, Perplexity, Copilot and Google AI Overview — across multiple regions. We monitor how those models answer and how they change over time, so you can stay ahead of shifts in the AI landscape.