Stereotypic behaviours (SBs) are linked with behavioural inflexibility and resemble symptoms of autism, suggesting that stereotypic animals could have autistic-like social impairments. SBs are also common in caged mice. We therefore hypothesised relationships between stereotypic and social behaviours, predicting that highly stereotypic mice would give/receive more agonism and be less effective in social learning tasks. Experiment One used C57BL/6 and DBA/2 mice in non-enriched or enriched housing (15 cages each); Experiment Two, more cages (six non-enriched, 44 enriched) plus a third strain (BALB/c). Across both experiments, enrichment reduced SB and agonism (aggression, plus ‘displacements’ where one mouse supplants another at a resource). These effects appeared related: housing effects on agonism became negligible when SB was statistically controlled for, and, at least in enriched cages, SB covaried with receiving aggression. In Experiment Three, 20 DBAs varying in SB from Experiment Two acted as demonstrators in a ‘social transmission of food preferences’ task. They were fed a novel flavour (shatavari powder), then each mingled with a familiar but flavour-naïve C57 observer. Observers were subsequently offered two novel flavours: shatavari or marjoram. Those spontaneously choosing more shatavari (n = 10) tended to have had less stereotypic demonstrators than the other ten observer mice. Overall, highly stereotypic mice thus received more agonism — an effect with obvious welfare implications that can be reduced with enrichment — and seemed potentially less effective at inducing flavour preferences in conspecifics. Such effects are consistent with social impairment, suggesting that reducing SB may perhaps enhance interactions between conspecifics.