In the context of the ongoing climate and biodiversity crises, mixed forest stands are increasingly considered as a sustainable management alternative to monocultures. We developed a new individual-based and process-based forest growth model, PDG-Arena, to simulate mixed forest functioning and test ecophysiological interactions among trees in mixed stands. The model builds upon the validated ecophysiological stand-scale model CASTANEA and integrates tree competition for light and water. We evaluated the performance of PDG-Arena by comparing the simulated growth with annual dendrochronological growth data from 37 common beech and silver fir monospecific and mixed plots in the French Alps. PDG-Arena showed a slightly better performance than CASTANEA when simulating even-age and monospecific forests (r2 of 32.1 versus 29.5%). When using structure-diverse and species-diverse inventories, PDG-Arena performed better than CASTANEA in pure beech (38.3 versus 22.9%) and mixed stands (40.5 versus 36.3%), but not in pure fir stands (39.8 versus 42.0%). The new model also showed a significant positive effect of species mixing on gross primary production (+5.5%), canopy absorbance (+11.1%) and transpiration (+15.8%). Our results thus show that tree-level process-based models such as PDG-Arena, formally simulating interspecific interactions, can serve as a valuable tool to understand and simulate the carbon, light and water dynamics of mixed stands.