Recent reports have shown that human immunodeficiency virus type 1 (HIV-1) Gag can directly affect susceptibility to protease inhibitors (PIs) in the absence of known resistance mutations in protease. Inclusion of co-evolved Gag alongside protease in phenotypic drug susceptibility assays can alter PI susceptibility in comparison with protease with a WT Gag. Using a single-replication-cycle assay encompassing full-length Gag together with protease we demonstrated significant variation in PI susceptibility between a number of PI-naïve subtype B viruses. Six publicly available subtype B molecular clones, namely HXB2, NL4-3, SF2, YU2, JRFL and 89.6, displayed up to nine-fold reduced PI susceptibility in comparison with the assay reference strain. For two molecular clones, YU2 and JRFL, Gag contributed solely to the observed reduction in susceptibility, with the N-terminal region of Gag contributing significantly. Gag and protease from treatment-naïve, patient-derived viruses also demonstrated significant variation in susceptibility, with up to a 17-fold reduction to atazanavir in comparison with the assay reference strain. In contrast to the molecular clones, protease was the main determinant of the reduced susceptibility. Common polymorphisms in protease, including I13V, L63P and A71T, were shown to contribute to this reduction in PI susceptibility, in the absence of major resistance mutations. This study demonstrated significant variation in PI susceptibility of treatment-naïve patient viruses, and provided further evidence of the independent role of Gag, the protease substrate and in particular the N-terminus of Gag in PI susceptibility. It also highlighted the importance of considering co-evolved Gag and protease when assessing PI susceptibility.