They were all built yesterday on v2.00.335....
I was building with aprox 1:1 stabilizer:reactor ratio, taking reactor block count divided by 8 then inverse cube and get the nearest cuboid to those dimensions, rounding up to odd.
Place your reactor, turn your y-axis angle to 45', set all your plains on, and place one group of stabilizers down so that the plains mirror all eight stabilizer groups with a gap of 3 between each group and at a distance from the reactor so as you have 100% stability. At that point it should be showing you 4/6 dimensions used.
From there just remove/replace the stabilizer groups increasing one of the gaps between them by one each time. (Basically just moving the stabs out in one dimension by one block at a time.) Each time you do this shift the other reactor alignment sliders, one at a time to 45'. To start with this will result in your dimensions falling from 4/6 to 2/6 and getting worse overall stabilization. But keep doing it and there will come a point where the dimensions go from 4/6 to 6/6. The stabilization will be way over 100% so you can now move the stabilizer groups back toward the reactor to reduce the stabilizer buffer as desired.... and Bob's your uncle. You've (ab)used the reactor alignment mechanic to achieve a 6 dimensional reactor in a configuration that results in the smallest possible total dimensional volume of the power system... which happens to be a long rectangular box.
ps. I found the easiest reactor size to demonstrate /test this with was the 199 reactor. the stabilizers can be nice 3x3x3 blocks and you don't have to do much to get to the 6/6 break point. Ie: reactor 199 @ 5x5x8(-1) and 8x stabilizer groups of 3x3x3 for a ratio of 199:216