English-speaking autistic children use the hesitation marker “um” less often than non-autistic children while they use “uh” at a similar rate. It is unclear why this is the case. We examined hesitation marker usage of Dutch-speaking autistic and non-autistic preschoolers with the aim to 1) make a crosslinguistic comparison of hesitation marker usage and 2) examine two hypotheses regarding the underlying linguistic mechanisms of hesitation markers: the symptom hypothesis and the signal hypothesis. We replicated group differences in “um” usage and showed significant effects of age and biological sex in the selection of hesitation marker (“uh” versus “um”). We found that “uh” usage is correlated with structural language abilities while “um” is not. “Um” may be related to pragmatic language abilities, which were not examined here. The findings support the notion that hesitation markers are not involuntary utterances but deliberate language features, which is in line with the signal hypothesis.