OpenAI has upgraded its custom GPT feature, allowing users to choose from all available AI models when creating or running a custom GPT. Previously, users could only build custom GPTs using GPT-4o.
Custom GPTs let users create personalized versions of ChatGPT with saved instructions, settings, and even files. This helps people avoid retyping prompts and speeds up tasks. For example, some users have made custom GPTs for studying, work research, or social media help.
Now, users can pick from a wide range of models like GPT-4o, GPT-4.5, GPT-4.1, GPT-4-mini, and others. There is also a “Recommended Model” option that sets a default model but allows users to switch models as needed.
However, there are two catches. First, if a custom GPT uses Actions (such as API calls, web browsing, or code execution), it can only run on a limited number of models. OpenAI says it plans to remove this restriction soon.
Second, this update is only available to Pro, Plus, and Team subscribers. Enterprise and EDU users will get access later.
This is the biggest update to custom GPTs since their launch in late 2023. Whether it will make more people use custom GPTs remains to be seen.
