Introduction: Negative symptoms are a common occurrence in patients with psychosis spectrum disorders. Previous analysis of the latent structure of the Clinical Assessment Interview for Negative Symptoms (CAINS)-which was developed to advance the assessment of negative symptomatology-showed two underlying sub-domains (Motivation and Pleasure; Expression). Recent findings indicate that a more complex structure might be more applicable. Aim: To evaluate the psychometric properties of the Serbian version of the CAINS in a sample of outpatients (N = 67) with psychosis spectrum disorders. Materials and Methods: Negative symptoms and general level of psychopathology were assessed with Serbian translations of the CAINS, the 53-item version of the Brief Symptom Inventory (BSI), and the 24-item version of the Brief Psychiatric Rating Scale (BPRS). Principal component analysis (PCA) was carried out on the CAINS items, and correlation analyses were done to assess its convergent and discriminant validity. Results: Our results showed an excellent internal consistency (Cronbach's alpha = 0.92). PCA revealed a three-component solution consisting of Expressiveness and Motivation for Social and Family Relationships (Factor 1), Motivation for Vocational Activities (Factor 2), and Motivation for Recreation (Factor 3). Convergent validity was supported by significant correlations with the Negative symptoms domain of the BPRS (Factor 1, 0.695, p < 0.01; Factor 2, 0.352, p < 0.05; Factor 3, 0.452, p < 0.01). When assessing discriminant validity, weak correlations were found with BPRS and BSI scores. Conclusion: The Serbian version of CAINS is a valid, reliable and useful tool for the assessment of negative symptomatology. Our findings support a three-factor structure of CAINS, which indicates that the construct is more complex than envisaged by the original conceptualization of two distinct factors.