This study examined the reliability and validity of the Virtual Assessment of Mentalising Ability (VAMA). The VAMA consists of 12 video clips depicting a social drama imposed within an interactive virtual environment with questions assessing the mental states of virtual friends. Response options capture the continuum of ability (i.e., impaired, reduced, accurate, and hypermentalising) within first- and second-order cognitive and affective theory of mind (ToM). Sixty-two healthy participants were administered the VAMA, three other ToM measures, and additional measures of neurocognitive abilities and social functioning. The VAMA had sound internal consistency and high test-retest reliability. Significant correlations between performance on the VAMA and other ToM measures provided preliminary evidence of convergent validity. Small to moderate correlations were observed between performance on the VAMA and neurocognitive tasks. Further, the VAMA was found to correlate significantly with indices of social functioning and was rated as more immersive, more reflective of everyday ToM processes, and was afforded a higher recommendation than an existing computer-based ToM task. These results provide potential evidence that the VAMA is an ecologically valid tool that is sensitive to the spread of ability that can occur in ToM subprocesses and may be a valuable addition to existing ToM measures. Future research should explore the validity and utility of the VAMA in larger, more diverse samples of healthy adult and clinical populations.