Recommended choice is in bold face:
For the other 1st row elements three pseudopotential versions exist. For most purposes the standard versions should be used. They work for cutoffs between 325 and 400 eV, where 370-400 eV are required to accurately predict vibrational properties, but binding geometries and energy differences are well reproduced at 325 eV. The typical bond length errors for first row dimers (N, CO, O) are about 1% (compared to more accurate DFT calculations, not experiment). The hard pseudopotentials _h give results that are essentially identical to the best DFT calculations presently available (FLAPW, or Gaussian with huge basis sets). The soft potentials are optimised to work around 250-280 eV. They yield very reliable description for most oxides, such as VO, TiO, CeO, but fail to describe some structural details in zeolites (i.e. cell parameters, and volume).
For HF and hybrid tpye calculations, we strictly recommend the use of the standard or of the hard potentials. For instance, the potential can cause unacceptably large error even in transition metal oxides, even though the potential works reliable on the PBE level.