Purpose: Autistic children with limited spoken language ability (LSLA) often do not respond to traditional interventions, reducing their social inclusion. It is essential to identify effective interventions, and sensitive measures to track their intervention response. Methods: Using data from an RCT comparing music-mediated and play-based interventions, we investigated the impact of spoken language ability on outcomes, and measured response to intervention through natural language sample measures. Results: Children with lower verbal IQ, relative to higher verbal IQ, made some greater gains over the course of music-mediated intervention. Natural language samples were helpful in characterizing communication and tracking change. Conclusion: Music-mediated interventions hold promise as effective interventions for autistic children with LSLA. Natural language samples are robust in characterizing this subgroup.