Although several studies have dealt with the use of derogatory terms on social media, only few compared the phenomenon across languages from a sociocultural aspect. This study used a mixed-method comparative analysis of 920 Arabic and English abusive tweets. The researchers used content analysis to annotate the tweets according to their type and severity. They also used qualitative thematic discourse analysis to interpret the linguistic themes. Furthermore, they used frequency analysis to statistically identify the most common targets and lexical items and to identify the sociolinguistic patterns behind them. The results reveal that Arabic tweets have higher frequencies of gender abusive terms, and they are more severe than the English ones. However, English showed greater reliance on vulgar terms because of cultural taboos. English communication was also dominated by implicit insults, while Arabic favored explicit offense in accordance with direct/indirect cultural values. Both languages used emojis intensively, but Arabic used more diverse registers within messages. Anonymity boosted prejudices for both languages. In conclusion, the difference in online toxicity between the languages is the result of linguistic differences and the cultural norms and the interaction between the two.