Final Magnetic Moment on site and database

I cannot understand correlation between Final Magnetic Moment given on site and in database.

E.g. for MnSi (mp-1431) data on the site is Final Magnetic Moment 0.005 \mu_B.
I tryed to reproduce it using downloaded VASP files and appropriate POTCAR and get the value 4.046 for magnetization (x) in column “tot”.
Database returns:

structure = mpr.get_structure_by_material_id(‘mp-1431’)
[1.023, 1.027, 0.994, 1.029, -0.048, -0.048, -0.049, -0.046]

That is, it is about 4 also. Even per formula unit it is 1\mu_B (as there are 4 f.u. per u.c.).

On the other hand, for Fe it is OK: 2.222 from site and similar from database:

structure = mpr.get_structure_by_material_id(‘mp-13’)
My repeated calculations gives 2.222 per atom.

Is there some weird thing not known for me, or is it a bug?

Thanks for reporting @ged, will investigate! The value reported by the MPRester definitely seems to be the correct value.