This research paper aims to compare machine learning techniques for detecting transient instability in power systems. The delayed construction of infrastructure and the increasing presence of converter-based power plants, which reduce system inertia, contribute to operating states that can lead to instability. In such scenarios, disturbances in the network backbone can disrupt system synchronism, potentially causing angle instability and leading to blackouts. Transient stability, characterized by rapid responses before a blackout, requires proactive defense schemes to predict the outcomes of disturbances and take preventive actions. However, predicting whether a disturbance will lead to instability or not is challenging due to multiple factors involved. Machine learning techniques offer the ability to address this challenge by leveraging their capacity for problem generalization. The results shows that several machine learning algorithms have very good accuracy to predict transient instability.