I've got a Grizzly G0463 benchtop mill, and I get my money's work out of it. It's a reasonably capable mill, within the size and rigidity limitations. I've had it mounted on several different workbenches in the four years I've owned it.
Workbenches through history
The first workbench I mounted the mill to was an absolutely beastly one. It was twelve feet long, three feet deep, and mounted permanently to the walls. My roommate/landlord at the time designed and built the workbench, and I got one end.
The second workbench I mounted my mill to was much smaller. It wasn't attached to any walls, so it was theoretically portable. In practice, it was still quite stationary. It was a bit too large to be useful. The mill only requires so much depth in a workbench, and any more means either the mill is set back too far, or there's a huge dead space behind the mill. This workbench was way too deep, and thus had a huge amount of wasted space.
I moved again, and decided to leave the second workbench behind (I'm sure my landlord was touched by my generosity). The third workbench I designed for the mill never made it past the design stage. It was going to be built from aluminum extrusion, but it would have been ridiculously expensive. I helped to make it expensive by designing it for less than .1" static deflection, with a 3:1 design factor (I had suspenders and a belt for much of the deflection calculations, but this thing was preposterously stiff). Designing for deflection would be a smart move if I were designing around resonant frequencies (the square root of static deflection is equal to the natural frequency), but I wasn't.
The latest iteration of the mill workbench is portable (on casters), reasonably cheap (constructed from lumber, not 80/20), and right-sized for the mill. There's not a ton of extra space, which is somewhat by design.