Abstract: As observational precision improves, a leading uncertainty in exoplanet interior inference is no longer the data but our models of realistic, multicomponent interiors. I present a mixing framework focused on planets less than 10 Earth-masses that treats differentiated layers as chemically consistent mixtures rather than endmembers. We implement this framework in the open-source planet interior code Magrathea. We solve the mineralogy of rocky mantles in closed form from four elemental ratios, Ca/Mg, Si/Mg, Al/Mg, Fe/Mg, using stoichiometric relationships to keep the mantle in chemical equilibrium across upper and lower regions. The resulting phase proportions of Fe–Mg silicates and accessory Ca- and Al-bearing minerals are broadly consistent with mineralogy from Perple_X Gibbs free-energy minimization but are obtained with a much faster and auditable scheme. We then use ideal mixing to translate these assemblages into density and an adiabatic temperature gradient. We extend the same machinery to tests of light elements in metallic cores, rock–water mixing, and mixed atmosphere species. In each case, we highlight where composition inferences are most sensitive to mixing assumptions. These developments build on the modular equation-of-state and phase-diagram infrastructure of Magrathea v2, enabling reproducible model intercomparisons where differences can be traced to specific materials, phases, and mixing choices rather than hidden defaults.