This work compares different linear and nonlinear RF energy harvesting models, including limited or unlimited sensitivity, for simultaneous wireless information and power transfer (SWIPT). The probability of successful SWIPT reception under a family of RF harvesting models is rigorously quantified, using state-of-the-art rectifiers in the context of commercial RFIDs. A significant portion of SWIPT literature uses oversimplified models that do not account for limited sensitivity or nonlinearity of the underlying harvesting circuitry. This work demonstrates that communications signals are not always appropriate for simultaneous energy transfer and concludes that for practical SWIPT studies, the inherent non-ideal characteristics of the harvester should be carefully taken into account; specific harvester's modeling methodology is also offered.