Yeah, never concrete examples from these guys.

I am creating a game and I can say that with the coding part the models help a lot, mostly gpt 5.5 high. Tbh to me all the frontier models feel the same and they can all solve the stuff I do quite well with some guidance and prompting. But that kind of makes me appreciate the other stuff more like visual style, sound design, mechanics etc etc. Tons of work still.

For brainstorming I find the models bad nowadays or maybe I am just too critical of the results