ChatGPT Free Users Now Default to Cheap GPT-5.2 Instant, No Auto Switch on Thinking
OpenAI has removed ChatGPT’s automatic switching feature for free ChatGPT Go users. Now, the AI chatbot will only use the cheaper GPT-5.2 instant model by default. But you can still choose to use the Thinking Model manually.
OpenAI has transformed the way ChatGPT and ChatGPT Go users work for free. Now, if you use ChatGPT for free or are a Go customer, the AI will direct you to the GPT-5.2 instant model, and will automatically switch to advanced reasoning with thinking models.
What has changed for ChatGPT free users?
In a blog post for the ChatGPT release notes, OpenAI said it is removing the automatic model router for free and ChatGPT Go users. The router, introduced only four months ago, was designed to meet user needs with quick responses from cheaper models or more thorough answers from thinkier models.
But users can still manually select the thinking model if necessary. “Free (and Go) users will now use GPT-5.2 Instant by default, and can choose to use Reason at any time by selecting Instant from the Tools menu in the message composer,” the release says.
This change takes effect with the launch of GPT-5.2 on December 11, 2025
The automated router will remain in place for ChatGPT Plus and ChatGPT Pro users.
Why did OpenAI make this change to ChatGPT?
According to a report by WIRED, automated routers led to greater adoption of the advanced logic model by free users, which increased from less than 1 percent to 7 percent. This increased operating costs as the thinking model performed more calculations for answers, and also led to longer wait times.
According to OpenAI, many users preferred quick responses over in-depth replies with longer wait times, the report said. Whenever users want more in-depth answers they have the option to choose a thinking model.
The change comes as OpenAI faces increasing competition from Google and its Gemini AI, with recent reports showing that ChatGPT’s user engagement has declined, and Gemini’s popularity increasing. Amid growing pressure in the AI field, OpenAI CEO Sam Altman has reportedly issued a ‘code red’ within the company.
OpenAI recently made a major improvement to ChatGPIT images, and also signed a major deal with Disney to use Marvel, Star Wars and other famous characters for its Sora video-generation model.
