I'm not directly involved in either project beyond reporting bugs and suggesting features yet, but I follow both projects closely. My sense is that the Mbin community is prioritizing collaboration around UX improvements while Kbin is focusing on scaling/performance issues... which makes sense as kbin.social is more than 10x the size of fedia.io (https://fedidb.org/network/instance/kbin.social vs https://fedidb.org/network). I opened a bug about the UI for altering link images at https://codeberg.org/Kbin/kbin-core/issues/1365. When I tested the same steps in Mbin, the issue i was seeing in Kbin had already been solved in Mbin.
Kbin is a great PHP implementation of ActivityPub for reddit-like communities, but requiring all major changes to be made/reviewed by a single person is a real bottle neck.
It would be great if Kbin could figure out some form of goverance/delegation that would allow more contributors, but there doesn't seem to be much interest in that type of change so for now we have 2 project with different priorities and governance models... and that isn't necessarily a bad thing.