Plant-based, i.e. vegetarian (without meat and fish) or vegan (exclusively plant-based foods) diets are in trend also among children and adolescents, but data on food intake in this group are lacking. Here, we compare the consumption of food groups of vegetarian (n 145), vegan (n 110) and omnivore (n 135) children and adolescents (6–18 years) in Germany using data of the VeChi Youth Study. Each food item reported in 3 d weighed dietary records was assigned to one of eighteen food groups and individual mean intake per day (g/MJ) was calculated. Group differences were assessed using covariance analyses adjusted for age, sex and other covariates. For food groups with a high number of non-consumers, non-parametric Kruskal–Wallis tests were run. Pairwise comparison of vegetarian and vegan groups indicated significantly higher intakes of legumes, nuts, milk alternatives (all P = 0·0003) and meat alternatives (P = 0·0065) among vegan subjects. Intake of these food groups of omnivore participants was low (Q3:0·0 g/MJ for legumes, milk alternatives and meat alternatives, 0·5 g/MJ for nuts). Dairy intake of vegetarians (11·6 g/MJ) was significantly lower than of omnivore subjects (24·7 g/MJ) (P = 0·0003). Intake of fats/oils and sweet foods was lowest in vegan compared with vegetarian and omnivore participants (P< 0·05). Whole grain intake was higher in vegan participants (14·5 g/MJ) than of vegetarian (9·1 g/MJ) and omnivore (6·5 g/MJ) participants (P = 0·0003). Longitudinal studies are necessary to evaluate the long-term health consequences of vegetarian, vegan and omnivore food patterns, especially in childhood and adolescence.