Using the RAevol model we investigate whether the molecular complexity of evolving organisms is linked to the " complexity " of their environment. Here, the complexity is considered as the number of different states environments can have. Results strikingly show that the number of genes acquired by an organism during its evolution does not increase when the number of states of the environment increases but that the connectivity of their genetic regulation network actually does. On the opposite, we show that the mutation rate has an important influence on the gene content. We interpret these results as a complex intertwining of direct selective pressures (the more genes, the better the organisms can be) and robust-ness and drift thresholds that limit the maximum number of genes at different values depending on the mutation rates.