Inconsistent formation energy for mp-1084832

Hi, I find that the formation energy per atom obtain for mp-1084832 is 0.4785 eV/atom when querying using MPRester, whereas it is -0.011 eV/atom according to the webpage (mp-1084832: Fe5Co3 (Cubic, Im-3m, 229)).
A manuel verification using the final energy obtained from MPRester shows that the latter should be correct.

Could you please confirm the issue? Can there be potentially similar issues for other materials?

Hi @mathsphy,

Are you using the old MPRester? (i.e. the one from pymatgen.ext.matproj)?

If so, you should switch to the new MP API:

The legacy API only has access to older database versions; some of the newer releases have fixed data inaccuracies in derived properties like formation energy.