Vapex (vapor extraction) is a nonthermal process that has significant potential in providing a more environmentally friendly and energy-efficient alternative to steam injection. Vaporized solvent injected in-situ dissolves into the oil and reduces oil viscosity, allowing the oil to flow to a horizontal production well via gravitational forces. While compositional simulators are available for assessing the Vapex performance, the simulation process may become difficult when taking into account the uncertainty due to reservoir heterogeneity. A semi-analytical proxy is proposed to model the process, in a way analogous to the steam-assisted gravity drainage (SAGD) model described by Butler, who demonstrated the similarity between two processes with a series of Hele-Shaw experiments and derived an analytical steady-state flow rate relationship that is comparable with the SAGD case. Solvent concentration and intrinsic diffusivity are introduced in this model instead of temperature and thermal diffusivity in SAGD. In this paper, analytical solutions and implementation details for the Vapex proxy are presented. The proposed approach is then applied to various reservoirs discretized with spatially varying rock porosity and permeability values; bitumen drainage rate and solvent penetration are calculated sequentially at grid blocks along the solvent–bitumen interface over incremental time steps. Results from this model are compared against experimental data available in the literature as well as detailed compositional simulation studies. Computational requirement of the proxy in comparison with numerical simulations is also emphasized. An important contribution from this work is that process physics are built directly into this proxy, giving it an advantage over other data-driven modeling approaches (e.g., regression). It can be used as an efficient alternative to expensive detailed flow simulations. It presents an important potential for assessing the uncertainty due to multiscale heterogeneity on effective mass transfer and the resulting recovery performance, as well as assisting decisions-making for future pilot and field development.