It is a bit baffling how people tend to say oh use GPT or Claude for anything AI related. Maybe this is just a Youtube tutorial thing but it sure stink of Apple users.
That makes sense to me if it were braindead Apple users who keep trying to shove GPT or Claude down my throat.
Personally I think people are sleeping on Gemini.
I guess I have not used GPT now for the longest time so I am actually not too sure how their interface has changes. I do know for a fact that interface excluding that GPT is incapable of sticking to your chosen model. That seems a pretty big irritant if you ask me.
As you can see in the screenshot Gemini out of the box also easily gives you plenty of options besides choosing other models. I personally don't fuck with the pro-anythings. I just have no need to unless it comes to code. Maybe a long running research thingy but generally, the same way even a person in a coma is great communication for me , 2.5 or what the latest model is usually works fine.
The reason why I like Gemini the most is because although it is Google it does not try to milk me in obvious ways. They also offer almost integrated everything. Aside from Stitch - Design with AI or Google AI Studio
I like how you can do almost anything from prototype a novel to plan out website in the chat itself. For everyday users such free features are super powerful... vs GPT crap.
It is also very usable on the phone which makes it accessible to many more people.
So all I am saying that sure if you are using the API's then waffle between different LLMs, but as an everyday AI , very few interfaces having anything on Gemini.
I'd really like it if you went into more detail about the specific models you use, because going forward it seems to be less and less about the AI type, and more about the models and what the next iteration offers?
For daily things I just use the free Gemini tier, which I think is currently Gemini 2.5 Flash. The type if you mean gemini vs gpt vs deepseek and Llama them still matters when you get to intensive things but yes for most people it is not a concern. If I were to use a model specific for coding then I would prefer to use Gemini 3 or GPT 5 , or if you are using models for iterative tasks but they are generic say like summarizing text then it can matter less how smart the model is. More so the model matters when you are doing local work with something like N8N , then you can use a smaller model with a small context window for simple tasks and use the bigger models with large context for producing things.
I don't analyze GPT or Gemini, but based on my past experience, I found GPT is better for me, but I guess the Gemini improved a lot as many people are also saying that it is doing well.
I think at the level of GPT and gemini and deepseek them for daily things it hardly matters which model you use. I just prefer the Gemini interface a lot more and they do seem to be adding more features at a better pace than GPT and I have never hit the free tier limit with Gemini either which I use to get with GPT all the time. Still, I guess it is just personal preference.