The present paper represents a part of a major scientific effort aiming to reveal possible effects, nutritional or health related, of genetically modified (GM) feed ingredients for Atlantic salmon. For 3 months groups of post-smolt Atlantic salmon were treated with diets holding 130 g kg1 of the protein as soybean, one which contained 800 g kg−1 GM type RR (Roundup Ready™), and compared with a standard counterpart (commercial hybrid, not isogenic line) analysed to be non-GM (nGM), and again compared with a standard fishmeal diet without soybean protein. All diets were composed to be within the category ‘compositional equivalent’ and held exactly the same proximate compositions, starch and sugar levels, above requirements for methionine and lysine, equal fatty acid profiles, vitamin, mineral and pigment contents. There was, however, a slight difference in levels of anti-nutrients between the three diets. The various dietary treatments resulted in more than tripling of fish weight in all groups. In addition no significant differences in feed utilization, whole body, liver and muscle proximate compositions, and no significant differences in muscle fatty acid profiles were measured. The diet without soybean of either type resulted in greater retention of lipid, but equal retention of protein (protein productive value). The relative sizes of liver, kidney, head–kidney and brain were the same in all dietary groups, while the relative size of the spleen showed significant differences between fish fed the genetically modified soy diet compared with fish fed the nGM soybeans. Fish fed the soy diets of either type also showed a somewhat reduced mean erythrocyte cell volume. All other haematological values were equal between diet groups. The detoxification system, measured as glutathione peroxidase (GPx) and lysozyme activities, showed equal values for all groups when measured in plasma and liver or head–kidney. The distal part of the intestine showed reduced sizes as an effect of soybean additions, without any differences between the GM and nGM type. Our results showed high growth, no mortality, haematological values within normal ranges, and efficient and equal responses in the detoxification system. This was a first indication that up to 130 g kg−1 RR-soybean protein can safely be used in diets for Atlantic salmon. However, there is still a need to elucidate higher inclusion levels of GM feed ingredients, and why spleen index was reduced, and if this was a long- or short-term effect.