- ChatGPT’s model picker has moved directly into the prompt box
- Instant is faster, while Thinking is better for complex tasks
- Pro gives subscribers ChatGPT’s most powerful option for harder workflows
OpenAI just quietly updated the ChatGPT interface for all users on Plus, Pro, and Business plans in a subtle way that makes it much easier to choose which version of its model you want to use — either Instant, for fast results, or Thinking, for a more considered response. But those aren’t your only choices, and there's a hidden setting, as we’ll explain.
Previously, these options were squirreled away in an obscure ‘ChatGPT’ menu at the top of the screen, which, frankly, not a lot of people knew about. Now they’re right where you write your prompt, and you can bet people will be using the different options a lot more.
The new menu leads to a choice between Instant, Thinking, and Configure. Here’s what you might not realize — the Instant option actually uses GPT-5.3, not the latest GPT-5.5. To use that, you need to select Thinking.
If you select Configure, you can specify exactly which model version you’d like to use, from older GPT models, going back to 5.4, 5.2, and o3.
When to use Thinking and Instant
ChatGPT-5.5 Thinking is better for hard tasks. It does a better job of keeping track of what it has already done. The way it delivers its results is different, too. It can sometimes show a short preamble before it starts reasoning. You can even add instructions while it’s ‘thinking out loud’ to adjust the final direction. You also get a much more streamlined output, with cleaner formatting and less unnecessary header text, even if it takes a long time to get your answer.
In contrast, ChatGPT-5.3 Instant is much better at delivering quick results. This can matter if you’re anything like me, because if I have to wait more than 10 seconds for something on the internet, I get distracted and start looking at something else, and before I know it, I’ve forgotten what I was actually doing in the first place.
So, if I had time and I wanted a more complete answer to a complex question like: “Given the current state of the market, how long do you think RAM prices will remain high?”, I’d use Thinking. If I wanted to know something simple, like “What is the average flight speed of a jumbo jet?”, I’d use Instant.
In the Configure menu, you’ll also notice that there’s an option for ChatGPT to auto-select which version of the model to use for you, based on the complexity of your question.
The auto-select switch is turned on by default, but there’s also the ethical element to consider. Instant uses less processing power in an AI data center, and therefore less water, which is used for cooling. It’s a bit like running your dishwasher using Eco mode. While it’s probably best for the planet not to have a dishwasher at all, at least running it in Eco mode uses up fewer natural resources.
You might think that there’s not much to the model version selector, but there is a hidden setting that’s worth knowing about.
If you move your cursor up to Thinking in the menu, then hover it over the option, you’ll notice that you get a choice between Standard and Extended. This refers to the length of time ChatGPT will spend thinking about your question or prompt.
So, if you want to extract the maximum amount of ChatGPT’s processing power, choose Extended here:
The last option in the Configure menu is Pro. This is the highest-capability GPT-5.5 option that ChatGPT currently offers. It’s for the hardest tasks and long-running workflows, and it is only available to Pro subscribers.
Just like Gemini
Of course, if you’re a ChatGPT user on a Free or Go plan, nothing will have changed in your web interface — you will still have the old ‘ChatGPT’ menu visible at the top of the screen. The interface in the app version of ChatGPT also remains unchanged. It’s unclear if this change will eventually roll out to the interface on these different plans and platforms in the future.
It hasn’t escaped my notice that the new look ChatGPT has gone for is exactly the way Gemini does it. It's interesting to see OpenAI changing part of its interface to be more like Gemini.
Now that a model option selector has appeared in exactly the same place in ChatGPT, instead of being hidden in an obscure location, you can bet it’s going to get used a lot more, and that will lead to better answers for everybody.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.
