For square-drive tools of different sizes but exactly similar geometry and identical materials, torsional strength theoretically should be proportional to the cube of nominal square size (or any other linear dimension). For example, ½" should be 2³=8 times as strong as a similar ¼" component, or (4/3)³=2.37 times as strong as an otherwise similar 3/8" component. The numbers quoted above don't follow that logic very well. Who knows why?
Wrenches are supposed to meet standards for strength and fit, for example U.S. Federal Specification GGG-WW-641E or ANSI B107.10M (which are both old, and may have been superseded by updated versions).
Perhaps the highest torque I ever put on a ½" drive handle was by standing on a flex-handle with a short pipe extension, maybe roughly 250 ft-ib., trying to loosen grossly overtightened lug nuts that were supposed to be 65 to 80 ft.-lb.