Abstract
Generative Artificial Intelligence (GenAI) has evolved into a transformative paradigm with capabilities spanning various modalities. Leveraging transformer-based architectures, diffusion models, and hybrid approaches, these systems have redefined the boundaries of creativity, productivity, and automation. This paper presents a structured survey of leading generative models, their system implementations, and diverse application domains. We categorize GenAI into six modalities which include text, image/video, speech, multimodal, music, and code generation, while also exploring system-level views. Finally, the paper highlights real-world applications underscoring GenAI’s growing role in reshaping human–computer interaction. The survey also briefly highlights some of the significant challenges, which are model bias, data requirements, and ethical issues, to give a balanced perspective on Generative AI systems.