OpenMusic provides the official PyTorch implementation of QA-MDT (Quality-Aware Masked Diffusion Transformer), a state-of-the-art text-to-music generation system. The model uses a masked diffusion transformer architecture trained on audio-text pairs to synthesize music from natural language prompts. It achieved top rankings on MusicCaps and Song Describer Dataset benchmarks and supports extended-length music generation in a zero-shot manner.