Social networks have a positive impact on everyday life, however, excessive and uncontrolled use can generate addictive pathological patterns. Objective. Propose a short version of the Social Network Addiction questionnaire (ARS-6). Method: 839 Peruvian adolescents and young people participate (M = 16.17; SD = 3.10), 58.8% women. A conceptual analysis of items based on the component model was implemented, different confirmatory factor analysis (AFC) models were identified to analyze the internal structure of the scale and measurement invariance analysis (AFC-MG). Results. The AFC models identified an overlap between the dimensions (correlations >.86) that was contrasted with a bifactor model, in addition, a non-parametric model of item response theory (Mokken) was implemented for scaling the short version. The invariance of measurement for the group according to gender was identified at all levels. Conclusion. The short version presents adequate evidence of the validity of its internal structure and reliability of the scores for its use.