Carbon‐bonded alumina foam filters are coated using special slurries based on graphene oxide (GO), carbon nanotubes (CNTs), and alumina nanosheets. The aim of such filters is to reduce the concentration of nonmetallic inclusions in the steel melt prior to casting into the molds. To assess the purification efficiency of the coated filters, a steel casting simulator is used. After immersion for a short time into steel melts with artificially generated alumina impurities, the filters are analyzed by optical microscopy and scanning electron microscopy. In addition, the population of remaining inclusions in the treated steel is investigated by the means of a special scanning electron microscope. The results show that, compared with a reference melt where no filter is used, the nanocoated filters cause inclusion growth and the formation of new silicate inclusions. The best purification efficiency, regarding only alumina inclusions, is given by the GO–C sample, whereas the combination of GO with alumina nanosheets or CNTs deliver worse performance.