Guns have always been a problem and the main cause of disruptions to public safety across the globe. This is a serious problem that should not be disregarded. All public areas can benefit from monitoring and surveillance services provided by an autonomous visual gun detection model. Gun detection has never been able to obtain the right speed or accuracy in real time in previous works. A reliable model to detect guns will enable a prompt response and suggest safety precautions. After investigating various research papers for a Yolo algorithms-based gun detection model, I used the Yolo algorithms with prediction heads with two different datasets. The dataset contains curated gun images that were collected from multiple sources to train and validate the Yolo models. The Yolo gun detection model is a dependable and effective model for various images of firearms and their orientations, achieving 87% precision and 70% recall. The SOTA (state of the art) aimed at the deep neural architectures for security purposes is approached by this detection model. The most recent Yolo-based gun detection model allows automated surveillance and alert systems to identify firearm threats more quickly in real time. This model’s performance is adequate for embedded applications and video in closed-circuit televisions. The primary difficulties arise in situations where there is inadequate lighting and the firearm is partially visible, making it challenging to identify the object. A smaller number of True-Negative and False-Positive cases have resulted from the experiments.