The war between cybersecurity specialists and cybercriminals (malware authors) continues to evolve. Both parties are trying to discover new, innovative and creative techniques. Machine learning and deep learning models have become the new battleground for both sides. This work focuses on malware detection and classification using machine learning, deep learning and Large Language Models between 2022 and 2024. The main contributions of this paper include the following: Firstly, it provides a comprehensive description of the machine learning and deep learning techniques recently used for malware detection and classification. Secondly, a comparison between classical classification models and new techniques such as the large language models. Thirdly, proposals to improve the efficiency of these techniques. This review aims to provide researchers with a clearer understanding of the field of malware detection, contribute to the advancement of cybersecurity and identify recent developments in the search directions that the scientific community is tackling to meet this challenge.