Firefly vs. GPT Image vs. Imagen

This bears some discussion on the future of Firefly, where it spectacularly fails on following a prompt while the other rendering models excel, like GPT Image and Imagen. I'm glad that there are alternatives Adobe has provided, but poor Firefly couldn't hurt a fly with its weak rendition of the prompt below.
"An ultra-high resolution of a 3D-rendered scene of an ultra-modern city, and on the left is a modern and tall and elaborately designed movie theater building with a large sign and large display movie posters covering the front windows, and people are going in and coming out , and on the right side is a busy supermarket with people going in and out of, and some are walking along the sidewalk. Ultra-fine detail, fine shading, beautiful contrasting lighting, on a partly sunny day. Shot from a street view looking down the street as the background fades out of focus."

The Firefly image on the left, then Imagen, then GPT Image