Social skills group training (SSGT) for children and adolescents with autism spectrum disorder (ASD) is widely applied, but effectiveness in real-world practice has not yet been properly evaluated. This 12-week pragmatic randomized controlled trial of SSGT compared to standard care alone was conducted at 13 child and adolescent psychiatry outpatient units in Sweden. Twelve sessions of manualized SSGT (KONTAKT) were delivered by regular clinical staff. Participants (N = 296; 88 females and 208 males) were children (n = 172) and adolescents (n = 124) aged 8 to 17 years with ASD without intellectual disability. Significant treatment effects on the primary outcome (Social Responsiveness Scale) were limited to parent ratings for the adolescent subgroup (posttreatment: -8.3; 95% CI = -14.2 to -1.9; p = .012, effect size [ES] = 0.32; follow-up: –8.6; 95% CI = –15.4 to –1.8 ; p = .015, ES = .33)and females (posttreatment: –8.9; 95% CI = –16.2 to –1.6; p = .019, ES = 0.40). Secondary outcomes indicated moderate effects on adaptive functioning and clinical severity. Conclusion: SSGT for children and adolescents with ASD in regular mental health services is feasible and safe. However, the modest and inconsistent effects underscore the importance of continued efforts to improve SSGT beyond current standards.