Increasing plant diversity in agricultural systems provides promising solutions for sustainably increasing crop yield. It remains unclear; however, how plant–plant interactions in diverse systems are mediated by plant genetic variation. We conducted a greenhouse experiment in which we grew three varieties of common beans with three companion plant species (chickpeas, sorghum, and sunflower) in different combinations (crop mixtures, bean cultivar mixtures, and monocultures), with and without drought stress. We hypothesized that under drought stress, the effect of companion plant species on bean yield would be mediated by the drought tolerance potential of the species. We further hypothesized that this effect would vary across different bean cultivars. Overall, we show that the effect of companion plant species on bean yield was not influenced by drought stress; instead, it was dependent on the identity of the bean variety. This could partially be explained by variation in growth rate between bean varieties, where the fastest growing variety recorded the highest yield increase in plant mixtures. The effect of companion plant species on chickpea biomass, however, was potentially influenced by chickpea drought tolerance potential; chickpea biomass was recorded to be higher in plant mixtures than in its monoculture under drought conditions. Our study highlights that to develop plant mixtures, it is not only important to consider the functional traits of the interacting plant species, but also those of the different plant varieties. We further suggest that stress tolerance can be a useful trait for initial selection of plant varieties when developing crop mixtures.