People tend to stick to one AI model. They got used to it, it worked for them once - so why change?
But when you work seriously with ChatGPT - especially when it comes to market research, marketing insights, and complex analyses - switching models isn’t a luxury or even the best practice. It’s a necessity.
Each model has its own style, strengths, weaknesses, and personality-as well as specific capabilities like web browsing, file support, or other features that Altman, in his infinite wisdom, decided to grant to one model and not another (may Altman’s name be blessed). The key is knowing when to use which.
Here’s how I work, for example, when doing market research:
I like to start with Deep Research, and not just with ChatGPT. You can use multiple sources that offer free research capabilities - GRNSPARK, ChatGPT, GROK, GEMINI - each brings up something different, surfaces a unique perspective, and more importantly: if one hallucinates, it’s easier to catch.
Then I compile everything into one document and let ChatGPT read it.
If I’m looking for deep insights, complex connections, or analysis of heavy CSV files - I turn to o1.
This model may feel less “marketing-oriented,” but for logic-heavy tasks, it’s more accurate, more methodical, and sees things that the more conversational models sometimes miss.
If I need the model to follow links, check up-to-date info, or simply have an open and easygoing conversation - I switch to o3 mini.
It may be less thorough than o1, but it knows how to handle the internet - and sometimes that makes all the difference.
And of course, if I want charts based on our conversation, the right move is to switch to GPT-4o and use its (relatively new) superpower: image generation.
The “logical” models - like o1 - might seem less suited for marketing and creative work at first glance, but they’re great at spotting patterns, testing assumptions, and extracting insights without getting distracted by pretty words.
So yes, I switch models all the time.
Not because it’s cool (even though it is), but because if I change the question - I also need to change who’s answering it.
Tag a friend who already switches models - and one who’s already being replaced by them. Just don’t tell them who’s who. 😏
