A model is proposed for the inverse miniemulsion polymerization of acrylamide (AAm) with an oil-soluble initiator. The model is based on the assumptions that (1) polymerization occurs in both the continuous phase and the polymer particles; (2) the radicals in the continuous phase with length that exceeds z crit can enter the monomer droplets; (3) a limited number of radicals can exist inside the polymer particles; (4) small radicals can exit the polymer particles; and (5) propagation is diffusion-controlled at high conversions. The model predictions of conversion, particle number, and the occurrence of a diffusion-controlled propagation reaction at high conversions were validated by comparison with experimental data for inverse miniemulsion polymerization, showing good agreement. The occurrence of diffusion-controlled propagation ("glass effect") at high conversions is also discussed and additionally validated with literature data.