During cooperative growth, microbes often experience higher fitness, due to sharing of resources by metabolic exchange and herd protection through biofilm structures. However, the trajectory of evolution of competitive species towards cooperation is not known. Moreover, existing models (based on optimisation of steady-state resources or fluxes) are often unable to explain the growth advantage for the cooperating species, even for simple reciprocally cross-feeding auxotrophic pairs. We present an abstracted model of cell growth that considers the stochastic burst-like gene expression of biosynthetic pathways of limiting biomass precursor metabolites, and directly connects their cellular levels to growth and division using a “metabolic sizer/adder” rule. Our model recapitulates Monod’s law and yields the experimentally observed right-skewed long-tailed distribution of cell doubling times. The model further predicts the growth effect of secretion and uptake of metabolites, by linking it to changes in the internal metabolite levels. The model also explains why auxotrophs may grow faster when provided the metabolite they cannot produce, and why a pair of reciprocally cross-feeding auxotrophs can grow faster than prototrophs. Overall, our framework allows us to predict the growth effect of metabolic interactions in microbial communities and also sets the stage to study the evolution of these interactions.ImportanceCooperative behaviours are highly prevalent in the wild, but we do not understand how it evolves. Metabolic flux models can demonstrate the viability of metabolic exchange as cooperative interactions, but steady-state growth models cannot explain why cooperators grow faster. We present a stochastic model that connects growth to the cell’s internal metabolite levels and quantifies the growth effect of metabolite exchange and auxotrophy. We show that a reduction in gene expression noise explains why cells that import metabolites or become auxotrophs can grow faster, and also why reciprocal cross-feeding of metabolites between complementary auxotrophs allow them to grow faster. Our framework can simulate the growth of interacting cells, which will enable us to understand the possible trajectories of the evolution of cooperation in silico.