Fishmeal (FM) was substituted with soybean meal (Glycine max) (SBM) and bambaranut meal (Voandzeia subterranea) (BNM) in 10 experimental African catfish, Clarias gariepinus, diets. Feed formulation was designed using mixture model. The inclusion level of the three protein ingredients varied between 0% and 60%. Remaining 40% comprised of basal ingredients kept constant for all 10 feeds. African catfish of average initial weight 35.2 ± 0.9 g were fed with one of the treatment diets for 28 days. The protein:lipid ratios of the diets (range 1.5–3.4:1) were used in evaluating the feed utilization and growth of the fish. We found that catfish performances were mainly depending on ingredients and not the ratio itself. The protein:lipid ratios in the diets made of plant ingredients were high but SGR was low. Specific growth rate (SGR) increased with the increase of feed FM content, being ca. 5% day−1 with 60% FM diet but ca. 2% day−1 at both 60% BNM and 60% SBM diets. SGR was similar (3.5% day−1) with diets of 30% BNM or SBM inclusion with 30% FM. Feed conversion ratio increased from below 0.6 of the 60% FM diet up to 1.5 (60% SBM) and 1.7 (60% BNM). Protein efficiency ratio decreased linearly with increasing FM substitution, but protein productive value (PPV) was similar for catfish fed 60% FM diet and its 50% substitution with BNM or SBM. These results suggest that protein:lipid ratio cannot be used in assessing nutritional performance if the source of feed ingredient vary widely. However, these results suggest that BNM can partly substitute FM and completely replace SBM in the diets of African catfish.