Aside from a few hiccups like this, I think it's looking good. To have 3D characters at all is rather impressive, especially considering the complications of showing them with custom clothing and equipment. I mean: I understand the point of having a small/short character float above a tall character in the same tile because that'd be one way that players could
at least see them. But it's just too silly-looking this way.
Suggestion 1: Instead of having multiple characters in the same tile standing one on top of each other, it would probably look a lot better to have them standing close to each other in close proximity to the tile area and each other. You could, for example, have one character in the center of the tile and the rest in quadrants around the center. Or you could have positions depend on the number of characters in the tile; i.e., if there were 2 then have them stand side by side, if there were 3 have them in a triangle formation, if there were 4... etc.
Suggestion 2: Maybe have the z-coordinate of the 3D model (height) be linked with the character's broadness trait? I'm just imagining that, for example, a dwarf could be depicted has being short and broad by using almost half the z-height as, say, a human, but using nearly identical x and y (width and depth?) as a human. And a skinny, slim elf could be depicted as slim by keeping the height close to norm, but with less width and depth.