On November 12, NetEase Cloud Music’s innovative generative recommendation model, named ‘Climber,’ was honored with the Best Application Paper Award at CIKM2025, one of the world’s most prestigious academic conferences. This cutting-edge model seamlessly integrates generative large-scale model technology with recommendation systems, comprehensively tackling the scalability issues inherent in the Transformer architecture. By adopting a multivariate hybrid sequence modeling strategy and a novel multi-task learning framework, ‘Climber’ not only achieves significant technological advancements but also offers the industry a new, directly applicable paradigm for refining recommendation systems. Presently, ‘Climber’ has been fully deployed across key scenarios within NetEase Cloud Music, including Crush Mode, Daily Recommendations, and Private Roaming, substantially elevating user satisfaction and operational efficiency. Furthermore, its architectural design has set a new industry standard, steering recommendation technology toward enhanced precision, efficiency, and personalization.