Previous prevalence rates of non-suicidal self-injury (NSSI) in adolescents have varied considerably. In the present cross-sectional study, prevalence rates, characteristics and functions of NSSI were assessed in a large randomized community sample consisting of 3,060 (50.5 % female) Swedish adolescents aged 15-17 years. The suggested criteria for NSSI disorder in the Diagnostic and Statistical Manual of Mental Disorders, 5th edition, (DSM-5) were used to assess prevalence rates with the aim of arriving at a more precise estimate. Out of the whole sample, 1,088 (35.6 %) adolescents (56.2 % female) reported at least one episode of NSSI during the last year, of which 205 (6.7 %) met suggested DSM-5 criteria for a potential NSSI disorder diagnosis. The NSSI disorder diagnosis was significantly more common in girls (11.1 % vs. 2.3 %, χ (2) (1, N = 3046) = 94.08, p < 0.001, cOR = 5.43, 95 % CI [3.73, 7.90]). The NSSI disorder group consisted of significantly more smokers and drug users compared to adolescents with NSSI that did not meet DSM-5 criteria for NSSI disorder, and also differed concerning demographic variables. A confirmatory factor analysis (CFA) was conducted on reported functions of NSSI, with the aim of validating Nock and Prinstein's (Journal of Consulting and Clinical Psychology 72:885-890, 2004, Journal of Abnormal Psychology 114:140-146, 2005) four-factor model on a Swedish community sample, resulting in a close to acceptable fit. A two-factor model (social and automatic reinforcement) resulted in a slightly better fit. The most frequently reported factors were positive and negative automatic reinforcement. A majority of functions were significantly more often reported by girls than boys. The implications of the suggested DSM-5 criteria and reported functions are discussed.