I was printing two phone models to hold them in my hand to compare. It's an embarrassing waste of plastic I know. But given that I wanted to make sure to use as little filament as possible.

I thought adaptive cubic would be the ticket but there was nothing "adaptive" about it. It put the same support on every layer.

Adaptive Cubic

I tried out lighting too and it behaved the same.

Lightning Infill

It did branch a bit but not enough to give me any confidence about top surface quality.

I ended up using height range modifiers. Starting with 5% adaptive cubic, a few layers of 10% adaptive cubic, then a few layers of 15% gyroid.

But I know if there were an infill which started low and gradually grew in density as it approached the top layer I could've save much more filament here.

Is there any slicer which has sort of a gradient infill that increases density as it approaches the top? Or is there a reason this is a bad idea? Or possibly adaptive cubic should've behaved differently here?