Visual speech substantially improves the recognition of spoken speech especially under noisy environmental conditions. For the recognition of monosyllabic words this benefit varies as a function of the relative strength of the auditory signal (signal to noise ratio: SNR) with a maximum benefit at “intermediate� SNRs, that is between conditions where the auditory signal is almost perfectly audible (high SNR) and where it is completely unintelligible (low SNR). The aim of this study was twofold: First, we asked whether this pattern is subject to change over the course of childhood development and second, we investigated whether it is atypical in children with autistic spectrum disorder (ASD). For that, we compared audiovisual gain over seven SNRs in typically developing children (TD) with that of neurotypical adults and a pilot sample of high functioning ASD children. Overall, TD children had more difficulty recognizing words when they were embedded in noise and did not experience as much gain from visual articulation as adults. This difference was only apparent at the lowest and the intermediate SNR. In adults, the point of maximal gain was located at a lower SNR than in TD children but was tied to a similar auditory- alone performance. ASD children experienced substantially less benefit than TD children while their auditory- alone and visual- alone performance remained unremarkable. Our data show that overall audiovisual gain increases throughout childhood development and is related to a decrease in susceptibility to noise in auditory speech recognition. This developmental pattern is reflected by an age- related shift of the maximal gain from higher to lower SNRs. Preliminary data suggest that ASD children experience less gain from visual speech.