Meet M6: The Chinese AI That Understands Text and Images at Scale

Researchers from Alibaba Group and Tsinghua University built M6, a 10B–100B parameter multimodal AI trained on 2TB of Chinese text and images. Using a Mixture-of-Experts architecture and large-scale distributed GPU training, M6 handles text, vision, and cross-modal tasks—generating images, answering visual questions, captioning products, writing poetry, and powering e-commerce AI at industrial scale.