Female children with autism spectrum disorder (FwASD) and performance intelligence quotient (PIQ) over 70 were compared with male children with ASD (MwASD) and typically developing (TD) controls (age 8–11 years) using a range of language and pragmatic measures. Functional ability was assessed using clinical observations and parent, teacher and self-reports. Results were compared between measures, and with direct assessments of language and pragmatics, in order to identify potential biases. This study found that FwASD performed better than MwASD but worse than TD controls on clinical observations of pragmatic ability. FwASD also performed worst overall on a parental measure of emotions. Additionally, there were patterns of differences between clinician, parent, teacher and self- reports and direct assessments, which indicate the need for assessment data to be collected from multiple informants. Findings also have implications for the accurate identification of ASD in females and appropriate provision of support.