SDXL vs SD1.5. No quality difference.

This was done on SD1.5, using a GTX980 card with 4GB of VRAM.

The SD1.5 "juggernaut_aftermath" model is smaller than the SDXL version of the Juggernaut model, but seems to work equally fine on 1024^2 (No doubles, as you can see there is just 1 person)

So is there really any reason to use SDXL, if smaller models on SD1.5 can produce similar quality? (When trained at 1024^2, like SDXL)

I can barely load the SDXL model of juggernaut (It's over 6GB, so with the Lowram settings in A1111 it really bogs the system down with RAM-swapping even though it does actually work despite my 4GB VRAM limit), while the SD1.5 version of juggernaut is much easier to handle (It's just around 2GB), and yet when comparing this image (Which is using a prompt posted on the SDXL page for juggernaut) to its SDXL counterpart it looks just as detailed and with similar level of quality (Fingers still look terrible on many SDXL images, but they look amazing on this SD1.5 image, so clearly SD1.5 is not automatically worse than SDXL)

(This image was generated without hires-fix)

The SDXL version of this image, by "mjtdev", is here: https://civitai.com/images/2724531?period=AllTime&sort=Newest&view=categories&modelVersionId=166909&modelId=133005&postId=641472 and in my opinion it's the exact same quality as the SD1.5 image I posted (Using the exact same prompt)