I’m a big fan of ChatGPT (so much so that I even enlisted its help to draft this message and check whether this feature has already been requested!). It looks like while there are discussions about reducing costs through model selection and fine-tuning, there isn’t a specific request for automatic model selection to balance cost and performance.
I’d love to see a setting that optimizes when the most advanced model (ChatGPT-4o) is used. Right now, it’s selected automatically, but many of my queries don’t require that level of power. A feature that prioritizes a lighter model for general tasks—while reserving 4o for complex problems—would help reduce costs and improve efficiency.
It would be great to have an option that either lets ChatGPT decide when 4o is truly necessary or allows users to toggle this behavior. Is this something OpenAI might consider adding?
I believe that User Control over which ChatGPTModel is used per message is important not only for reducing costs but also for enhancing the user experience.
Having the ability to control which model handles my messages — whether it’s the full GPT-5 for complex queries or a “mini” lighter model for casual or simple inputs like greetings or quick questions make it more personable.
I don’t want to worry about “wasting” a prompt or inquiry for something simple like a greeting or general search.
Additionally, I’m current on the free plan and it would help users like me:
Save premium message usage for impactful prompts.
Use the lighter model for quick chats without wasting quota.
Better tailor the AI interaction to different use-case scenarios.
It would be great if this choice could be offered as a toggle or command in the chat interface.
Ideally, this will further develop the AI’s machine learning when it comes to understanding use-case scenarios and there won’t be a need for a toggle switch down the road. I mainly dislike the fact that my first 10 messages entered are powered by GBT-5 whether I want them to be or not.