Needed to debug:
Seen on my MK4, with MMU3 and MMU plugin version 2025.8.16
Slicer shows the used filament in g to be significantly higher than the estimates that appear in octoprint and are subsequently passed on to Spoolman, I'm guessing it's not including the purged filament used in tool changes?