After my standard test run I have to say the Hillside performs pretty well on verbose abstract prompts - it's fairly creative and gives varying results (which is a plus for abstract verbose prompts). But otherwise I failed to discover the purpose of the model. Nothing is said in the description and in stylised images it performs less well than my favourite models. Maybe I just didn't discover its strengths (a description would be nice to have). Upskasling with this checkpoint up to x1.5 is hit-and-miss and upscaling to x2 wrecks the image. Overall it works kinda similarly to Juggernaut, which isn't bad at all, but hard to justify using it since it's doesn't seem to do anything in particular better than the other popular models out there.

What I love about Hillside is the hair, very different from many other checkpoints.

It has 0 understanding of NSFW prompts, it draws only faces.