The LabSOCS software is one of the main simulation codes employed to calculate the full-energy peak efficiency (FEPE) for the measurements of radionuclides by gamma-ray spectrometry. The validity of LabSOCS has not previously been evaluated for a large variety of matrixes, geometry types, and energies. This study aims to analyze in depth the LabSOCS FEPE robustness comparing the LabSOCS and experimental FEPEs by using three coaxial Ge detectors for a wide range of samples, varying the matrix composition, counting geometry (thickness (h) for cylindrical geometry), gamma emission energy (Eγ), and bulk density (ρ). The FEPE averages and variances calculated for numerous matrix types prepared using gamma cocktails were statistically comparable for a wide Eγ range (from 46 keV (210Pb) to 1836 keV (88Y)). To test the accuracy of the LabSOCS FEPEs varying h (6–50 mm), four certified samples containing only natural radionuclides were used. The 234, 228Th, 228, 226Ra, 210Pb and 40K activities were determined for each h, obtaining |zscore| < 2 in general. This also proves the good reproducibility of LabSOCS calculating activity concentrations. The LabSOCS FEPE robustness was also tested for a wide ρ range (1.3–2.8 g cm−3) mixing a sediment sample with Pb(NO3)2 for h = 50 mm. The 210Pb and 234Th contained in each mixture were determined, achieving all |zscore| < 2. Furthermore, the ability of LabSOCS to correct self-attenuation effects was checked to be wider than when other correction models are applied.