There's some really impressive things about this (the speed, the lack of typical AI image gen artifacts) but it also seems less creative than other models I've tried?

"mountain dew themed pokemon" is the first search prompt I always try with new image models and Nano Banna Pro just gave me a green pikachu.

Other models do a much better job of creating something new.

IMHO I'd rather them focus on strong literal prompt adherence so that more detailed prompts produce more accurate results.

That way you can stick your choice of any number of LLM preprocessors in front of a generic prompt like "mountain dew themed pokemon" and push the responsibility of creating a more detailed prompt upstream.

https://imgur.com/a/s5zfxS5

Note: I'm not particularly impressed with either of the results - this is more a demonstration.