Introduction Investigations of patterns of sexual arousal to certain groups of sexually explicit media (SEM) in the general population in non-laboratory settings are rare. Such knowledge could be important to understand more about the relative specificity of sexual arousal in different SEM users. Aims (i) To investigate whether sexual arousal to non-mainstream vs mainstream SEM contents could be categorized across gender and sexual orientation, (ii) to compare levels of SEM-induced sexual arousal, sexual satisfaction, and self-evaluated sexual interests and fantasies between non-mainstream and mainstream SEM groups, and (iii) to explore the validity and predictive accuracy of the Non-Mainstream Pornography Arousal Scale (NPAS). Methods Online cross-sectional survey of 2,035 regular SEM users in Croatia. Main Outcomes Measures Patterns of sexual arousal to 27 different SEM themes, sexual satisfaction, and self-evaluations of sexual interests and sexual fantasies. Results Groups characterized by sexual arousal to non-mainstream SEM could be identified across gender and sexual orientation. These non-mainstream SEM groups reported more SEM use and higher average levels of sexual arousal across the 27 SEM themes assessed compared with mainstream SEM groups. Only few differences were found between non-mainstream and mainstream SEM groups in self-evaluative judgements of sexual interests, sexual fantasies, and sexual satisfaction. The internal validity and predictive accuracy of the NPAS was good across most user groups investigated. Conclusions The findings suggest that in classified non-mainstream SEM groups, patterns of sexual arousal might be less fixated and category specific than previously assumed. Further, these groups are not more judgmental of their SEM-related sexual arousal patterns than groups characterized by patterns of sexual arousal to more mainstream SEM content. Moreover, accurate identification of non-mainstream SEM group membership is generally possible across gender and sexual orientation using the NPAS.