Ensemble perception, the ability to rapidly and automatically assess the summary or ‘gist’ of large amounts of information presented in visual scenes, is available early in typical development (Sweeny et al., in press). Ensemble perception might be compromised in children with autism, who have been proposed to present limitations in their abilities to maintain and/or use summary statistics representations for the recent history of sensory input, which might be responsible for their unique perceptual experience (Pellicano & Burr, 2012). Here we examined ensemble perception of emotions in 29 children with autism, aged between 6 and 14 years, 30 age- and ability-matched typical children, and 12 typical adults. Participants received three child-friendly tasks: a) an average (ensemble perception) emotion discrimination task, assessing their ability to judge the average happiness of a set of ‘morphs’; b) a non-average happiness discrimination task, evaluating baseline emotion discrimination; and c) a face identification task, estimating children’s ability to identify morphs that had been previously presented to them. We also monitored participants' eye movements while they performed the three tasks using a Tobii X2-30 eye tracker (30 Hz). All three groups presented better precision in the non-average than in the average emotion discrimination task, while children’s precision and accuracy in the three tasks were worse than those of adults- suggestive of a gradual and parallel maturation of emotion discrimination, ensemble perception, and face-processing abilities. Unexpectedly, children with autism were indistinguishable from typical children in their precision and accuracy across the three tasks. On eye-movement variables, the three groups did not differ in terms of the average number of fixations and the number of morphs sampled per trial. Contrary Pellicano and Burr (2012), these findings suggest that autistic children’s abilities for ensemble perception of emotions are largely similar to those of typical children.