Natural language processing (NLP) is a highly captivating area of study that has generated a substantial amount of research in modern times. Implementing multilingual harmful comment classification in the algorithm can provide a valuable advantage in analyzing comments, tweets, and other social media content related to a recognized topic. This will enhance the comprehension and examination of the miscommunication that occurred on many social media platforms regarding a specific topic, as well as the deterrence of deceitful remarks. The process of classifying multilingual harmful comments entails analyzing a hypothetical sentence that is formed based on an assumption. The suggested hypothetical language can be classified into one of three categories: it can be neutral, contradicting the known premise statement, or implying the premise. Natural language inference is a prominent topic in the field of NLP. Its objective is to determine the connection between two statements based on a given premise and hypothesis. Therefore, the research presents a model that aims to forecast the correlation between two claims. The prediction aids in determining whether the supplied hypothesis is in an entailment, neutral, or conflicting relationship with the given premise.