Mochi 1, developed by Genmo, is a groundbreaking open-source AI model designed for high-quality video generation based on text prompts. Utilizing a 10 billion-parameter model, it enables smooth, lifelike videos with realistic motion dynamics, including fluid and hair simulations. Mochi 1 is powered by Genmo’s proprietary Asymmetric Diffusion Transformer (AsymmDiT) architecture, which enhances efficiency in processing text and visual cues. It generates 30 fps videos up to 5.4 seconds, providing developers, AI researchers, and creatives with a powerful tool for building engaging video content with fine control over details.