The gas condensate reservoirs, which are a major source of energy, are experiencing an unwanted phenomenon called liquid drop-out. Gas injection as a proven effective remedy is generally employed to diminish the liquid loss. The condensate recovery by gas injection is typically estimated by assuming local equilibrium. This assumption is not valid in porous media when the limited contact area, high velocity, and low residence time are involved, particularly in near wellbore region. In this study, we attempt to calculate mass transfer coefficient of the gas condensate system in porous media at the transient state. Hence, a sets of experiments are performed where the tests include three injection gases (CO2, N2, and CH4), three liquid types as condensate components (C5, C6, and C7), and three mean grain sizes (150–300??m) within a wide range of gas velocity (0.00472–0.0283?cm/s). The experiments are modeled numerically and analytically to fit the effluent concentrations. The obtained mass transfer coefficients are then employed to introduce new models such that Sherwood number (Sh) is correlated to Peclet (Pe) and Schmidt (Sc) numbers, and other physical/operational parameters. Based on systematic sensitivity analysis as well as the modeling results, both analytical and numerical approaches are promising and viable. The statistical parameters are calculated to be 97.13% (R2) and 0.0044 (MSE) for the analytical method, and 97.33% (R2) and 0.00303 (mean squared error or MSE) for numerical solution, confirming the accuracy and reliability of employed models. The mass transfer coefficients are then fitted well using three correlations with R2 values of 95.29%, 97.04%, and 96.73%, and absolute relative deviation or ARD magnitudes of 13.22%, 11.73%, and 11.44%. The new proposed correlations can be simply implemented in numerical/analytical modeling and commercial software packages to obtain the non-equilibrium mass transfer coefficients of components in the gas phase. The se