Extant studies have examined factors contributing to perception of attractiveness of individual human faces. Because those studies primarily focused on ratings of at- tractiveness of a single target face, it was unclear whether observers could perceive attractiveness of a group of people as a whole. The present study examined whether observers could compare the group-wide attractiveness between two groups consisted of multiple members. We predicted that observers should be able to discriminate which of the two groups was higher attractiveness. Observers were briefly (1500 ms) exposed with two frames of images each of which consisted of four faces and determined the one that they believed more attractive as a whole. The results indicated that discrimination accuracy was above chance level. Virtually identical pattern of the results was obtained when each group consisted of eight faces in Experiment 3 and when exposure duration was 500 ms or 100 ms in Experiment 4. These results suggest that observers could per- ceive attractiveness of a group of people as a whole when discriminating attractiveness of two groups of people.