Here is the block moi file imported into C4D as a lower res poly model shown in the 1st image.
The model was then cleaned up manually using slide points and optimization/welding points.
Then mirrored the 1 quarter part to make a whole.



Now in the second image the final model is hypernurbed by 1 iteration.
It was a quick edit, so didn't pay that much attention to making quads
Detail is still retained. Not sure why you would think it would melt it.....All in how you lay out the polys.



Admittedly for a complex model you might pull your hair out ;)