This work focuses on the application of nanofiltration (NF) to the concentration of a pharmaceutical product, Clavulanate ($CA^-$), from clarified fermentation broths, which show a complex composition with six main identified ions ($K^+$, $Cl^-$, ${NH_4}^+$, $H_2{PO_4}^-$, ${SO_4}^{2-}$ and $CA^-$), glucose and glycerol. The solutes transport through the NF membrane pores was investigated using the SEDE (Steric, Electric and Dielectric Exclusion) model. This model was applied to predict the rejection rates of the initial feed solution and the final concentrated solution (10-fold concentrated solution). The best results were achieved with a single fitted parameter, ${varepsilon}_p$ (the dielectric constant of the solution inside pores) and considering that the membrane selectivity is governed by steric, electric (Donnan) and Born dielectric exclusion mechanisms. While the predicted intrinsic rejections of solutions comprising up to six ions and uncharged solutes were in good agreement with the experimental values, the deviations were much larger for the 10-fold concentrated solution.