Three diets were formulated to contain 16% crude protein (CP), 12.8 MJ/kg digestible energy (DE) and 0.9% lysine on an air-dry basis. The diets either contained 8.3% fishmeal, 20% faba beans plus 7.9% soybean oilcake meal (SBOK) or 20% low alkaloid lupins plus 8.8% SBOK as protein sources. The latter two diets were compared in a metabolism and nitrogen (N) balance trial using a completely randomized design and six pigs/treatment. All three diets were evaluated in a growth trial using 10 pigs (five boars and five gilts) over the growth interval 30-90 kg. The experimental design was a 3 (diets) x 2 (sex) factorial. Carcass characteristics were determined at the end of the growth trial. Pigs in the metabolism and N balance trial consumed 5% less (P<â€“ 0.01) of the lupin diet compared to the faba bean diet (1 370 vs. 1 440 g/d). No significant differences in the digestible energy (DE) content and in apparent N retention between the two diets occurred. In the growth trial lower intake levels (P<â€“ 0.01) were once again observed in pigs on the lupin diet (2 489 g/d) compared to the faba bean (2 637 g/d) and fishmeal diets (2 668 g/d). Pigs on the lupin diet grew (P< 0.01) slower than pigs on the other two diets (768 g/d vs. 857 g/d and 869 g/d for the faba bean and fishmeal diets respectively). The feed conversion ratio of pigs on the lupin diet (3.27 kg/kg gain) tended to be poorer than for those on the faba bean (3.03 kg/kg gain) and fishmeal diets (3.12 kg/kg gain). No significant differences were observed in dressing percentage (mean value of 78%), eye muscle area (mean value of 41.3 cm2) or P2-back-fat thickness (mean value of 16 mm) of pigs on the different diets. Both faba bean and low alkaloid lupins can be used at levels of up to 20% in diets of growing pigs, although lower production figures may be expected on lupin diets.