A detailed understanding of the air–fuel mixing process in gasoline direct injection engines is necessary to avoid soot formation that might result from charge inhomogeneities or liquid fuel impingement on the cylinder walls. Within this context, the use of multidimensional models might be helpful to better understand how spray evolution in cylinder charge motions and combustion chamber design affects the mixture quality at spark-timing. In this work, the authors developed and applied a computational fluid dynamics methodology to simulate gas exchange and air–fuel mixture formation in gasoline direct injection engines. To this end, a suitable set of spray submodels was implemented into an open-source code to properly describe the evolution of gasoline jets emerging from multihole atomizers. Furthermore, the complete liquid film dynamics was also considered. For a proper assessment of the approach, a gasoline direct injection engine running at full load was simulated and effects of spray targeting and engine speed were studied. A detailed postprocessing of the computed data of liquid film mass, homogeneity index and equivalence ratio distributions was performed and correlated with experimental data of particulate emissions. Satisfactory results were achieved, proving the effectiveness of the proposed methodology in predicting the effects of injection system and operating conditions on soot formation.