Fitted functions that correlate binary interaction parameters (BIP) for the perturbed-chain statistical associating fluid theory (PC-SAFT) equation of state (EoS) with the carbon number of n-alkanes were established for gas solubility in n-alkanes using solubility data of gases in a few selected n-alkanes typically with 10-28 carbon atoms. Data screening was made to ensure accuracy and consistency of data used for parametrization and validation. The PC-SAFT EoS using the BIPs by the established fitted functions was evaluated against experimental data. It was found that the PC-SAFT EoS is quantitatively accurate for the entire homologous series of n-alkanes, regardless of whether the data of n-alkanes were used in the parametrization. It was shown that the PC-SAFT EoS successfully reproduces the two types of the temperature dependence of gas solubility: (1) gas solubility decreases with an increasing temperature and then increases; (2) gas solubility always increases with an increasing temperature. In particular, the PC-SAFT EoS is capable of predicting an experimentally observed linear relationship between gas solubility in n-alkanes and the carbon number of n-alkanes. To put it into perspective, the PC-SAFT EoS was compared to a couple of popular benchmark group-contribution (GC) EoSs, namely, the predictive Soave-Redlich-Kwang (SRK) EoS and the SRK EoS with the modified Huron-Vidal second-order mixing rule. It was found that the two models generally deliver much worse results than the PC-SAFT EoS despite their satisfactory performance for gas solubility in lighter n-alkanes. The PC-SAFT EoS and other cubic EoSs with improved accuracy in modeling of asymmetric mixtures compared to the above benchmark models were validated against experimental data of synthetic gas condensates. It was found that the PC-SAFT EoS is superior to the other models. The comparison was also made among different parametrization strategies, namely, GC-PC-SAFT, the one proposed in this study, and a four-parameter correlation for the simplified PC-SAFT EoS. It was shown that the PC-SAFT EoS parametrized in this study yields more accurate results, although the other two approaches also give satisfactory performance.