ChatGPT model picker explained: Instant, Medium, High, Extra High, Pro, and which model to use
Updated: Sep 17
The ChatGPT model picker in 2026 is mostly a reasoning-control interface, not a menu where every label corresponds to a different standalone base model.
Instant, Medium, High, Extra High, and Pro change how ChatGPT allocates capability, but the underlying model depends on the plan and the product surface.
··········
THE PICKER NOW CONTROLS REASONING DEPTH
Medium, High, and Extra High are not separate GPT-5.6 model families. They change the reasoning effort of GPT-5.6 Sol on eligible paid plans.
........
Picker choice | Underlying path | Best use |
|---|---|---|
Instant | Luna on Free/Go; Sol on eligible paid plans | Fast everyday work |
Medium | GPT-5.6 Sol | Balanced reasoning |
High | GPT-5.6 Sol | Harder analysis and coding |
Extra High | GPT-5.6 Sol on eligible higher plans | Maximum Sol reasoning |
Pro | GPT-5.6 Sol Pro or GPT-6 Pro where eligible | Hardest and longest-running tasks |
........
··········
INSTANT IS PLAN-DEPENDENT
Instant uses GPT-5.6 Luna on Free and Go and GPT-5.6 Sol on eligible paid plans. The same visible label can therefore sit on a different underlying model depending on the subscription.
··········
THINK ON FREE AND GO DOES NOT UNLOCK SOL
The Think control for Free and Go uses GPT-5.6 Luna. It increases reasoning effort within that access tier rather than silently moving the request to GPT-5.6 Sol.
··········
MEDIUM AND HIGH ARE THE MAIN PAID REASONING LEVELS
Plus and higher eligible plans can use Medium and High with GPT-5.6 Sol. These are appropriate when the task benefits from deeper planning, analysis, or coding but does not require a Pro-class model.
··········
EXTRA HIGH IS THE TOP SOL REASONING LEVEL
Extra High pushes GPT-5.6 Sol further on eligible plans. It remains Sol, so the difference is reasoning allocation rather than a new base-model identity.
··········
PRO CAN MOVE TO A DIFFERENT MODEL CLASS
The Pro option can use GPT-5.6 Sol Pro and, where the plan permits, GPT-6 Pro. These models have separate usage allowances and should be treated as a different capability tier rather than merely a slider position.
··········
ASKING CHATGPT TO THINK HARDER DOES NOT ALWAYS CHANGE THE PICKER
On Plus and Pro, OpenAI says Instant does not automatically switch to a higher thinking level simply because the prompt says 'think harder'. The user should select the desired reasoning level when the task needs it.
··········
A PRACTICAL MODEL-SELECTION RULE
Use Instant for routine questions, Medium for normal analytical work, High for difficult reasoning or coding, Extra High when Sol needs maximum depth, and Pro only when the task justifies the higher-capability model and tighter allowance.
··········
FOLLOW US FOR MORE
··········
DATA STUDIOS
··········
