This research report proposes a typology of regulatory models for combating online hate speech, an issue for both human rights and democracy. The typology shows that the regulatory response has evolved over time and in different territories. Initially, social media companies were seen as direct addressees of regulation (a model of American origin based on the idea that social media companies are exempt from liability). Later, they were seen as intermediaries between the State and users (a model of European origin based on the idea that social media companies should contribute to the fight against online hate speech). At a territorial level, the European model for regulating social media companies also has its nuances, with measures that have been progressively strengthened both from the point of view of relations between social media companies and public authorities and between social media companies and users.