We present a crowding-based model to predict the shear viscosity of suspensions of rigid bimodal-sized particles. In this model, the mutual crowding factor is defined to explicitly account for the change in the amount of fluid trapped in the interstices formed by particles upon mixing particles with two different sizes. Through this factor, we cancel the effect of size interference by mapping the bimodal suspensions to a suspension of noninterfering size ratio. This approach provides a set of decorrelated particle fractions that depend on crowding and size distribution. The shear viscosity of the resultant suspension is then directly estimated based on the viscosity of corrected components using a stiffening function that accounts for the self-crowding in each size class individually. We tested the proposed model against published experiments over a wide range of particle volume fractions, and we observe an excellent agreement.