Gas entrainment (GE) is a critical phenomenon affecting the performance and safety of sodium-cooled fast reactors (SFRs). However, accurate numerical prediction of GE remains challenging due to the complex interaction of multiscale flow structures and highly deformable gas–liquid interfaces. Recently, a two-phase Lattice Boltzmann Method (LBM) with double-distribution-function [1] was introduced to simulate GE. It offers higher accuracy than simplified vortex models and better efficiency than incompressible Navier–Stokes solvers. Nevertheless, achieving high-fidelity results requires very fine spatial resolution, making the large memory demand of the two-phase LBM a major limitation. To address this issue, this study develops a hybrid two-phase Lattice Boltzmann Method–Finite Volume Method (LBM–FVM) method that reduces memory requirements while maintaining accuracy and computational efficiency. In the proposed approach, a velocity-based LBM with a cumulant collision model is used to solve hydrodynamics, providing high numerical stability and accuracy. The resulting velocity field is interpolated at cell faces and supplied to a Finite Volume Method (FVM) solver for the phase-field Allen–Cahn equation. Second-order isotropic centered stencils are employed in diffusive and sharpening fluxes, a third-order TVD MUSCL scheme for advection, and a second-order TVD Runge–Kutta method for time integration. The updated phase-field variable is then used to construct density and viscosity contrasts for the subsequent LBM step. The hybrid method is validated against a quasi-steady vortex experiment [2], where vorticity in a water tank induces gas entrainment. The simulations successfully reproduce key entrainment mechanisms, including circulation, vortex core, and GE formation. Compared to conventional double-distribution-function LBM, the proposed method significantly reduces memory usage while maintaining accuracy and computational cost, enabling high-resolution simulations that closely match experimental results. This makes the method a promising tool for large-scale GE simulations.