Changes to automatic switching to thinking in ChatGPT (Plus and Pro)

What changed

Source summary

OpenAI is retiring automatic switching from Instant to Thinking mode for ChatGPT Plus and Pro users globally; going forward, users must manually select a reasoning/thinking option from the model picker rather than have ChatGPT decide when to switch.

Why it matters to your work

Practical interpretation

If you're on ChatGPT Plus or Pro, ChatGPT will no longer automatically switch itself into slower 'thinking' mode for harder questions — you'll need to manually pick a reasoning model from the model picker when you want deeper analysis, so it pays to know that option exists rather than assume ChatGPT will do it for you.

When this could help

If you relied on automatic reasoning for complex questions, you now need to manually select a thinking option in the model picker when you want more reasoning time.

OpenAI has retired automatic switching from Instant to Thinking mode for Plus and Pro users globally, and removed the 'Higher intelligence' web setting. You now choose when to use reasoning mode rather than having it applied automatically.

How to get started

1. When you need reasoning for a complex question, select a thinking option from the model picker instead of relying on automatic switching.

   Provider guidance ↗

Build it into your routine

SCALA suggested workflow

1. Conscious model selection

   Before asking a complex question that benefits from extended reasoning, manually switch to a thinking model. For quick factual lookups, stay on Instant to save time and usage.

What to check

• ChatGPT can still switch to reasoning automatically for safety reasons, even though the general automatic switching has been retired. Source ↗