In this paper, we explore three popular approaches: GANs, LSTMs, and VAEs, to explore the fascinating world of computer-generated music. To envision GANs, imagine two creative friends working together to create a musical composition; one works diligently on the composition while the other serves as the discerning judge, ensuring its aesthetic appeal. As a memory reservoir, LSTMs, on the other hand, preserve the essence of previous melodies to serve as an inspiration for and guide in the creation of new harmonious compositions. Like a musical blender, VAEs expertly blend song fragments to produce completely new and melodic musical compositions. Our main goal is to determine which of these techniques produces the most enjoyable musical result while also pointing out any potential drawbacks. We can learn a lot about the future of computer-generated music, and its transformative potential in reshaping our listening experiences and perspectives on music composition, by carefully examining these methodologies.