At this point a lot of people use "ChatGPT" to refer to the ChatGPT model with the GPT-3.5 backend

To me it's a collection of both, so the title makes it sound that it performs better than GPT-4. I don't know who those people are exactly who only hold GPT-3.5 to that meaning.

Probably the free users.

I imagine the paid users are a minority, so almost everyone

Even in Hackerrank and amongst technical people whom the article I would imagine is directed to?

And a lot of people don't. In particular, all the stories in the media about ChatGPT's unexpectedly good performance at various tasks, are more likely than not to be about GPT-4 model.

This is a legitimate point of confusion, and it's best to keep it explicit which model is being discussed. Really should be written at the start of the text.

Thanks! will replace "chatgpt" with "gpt-3.5-turbo".