MXMACA (MetaX Advanced Computing Architecture) is MetaX's unified heterogeneous computing platform, built on a dual-track strategy of "independent innovation and open compatibility". Leveraging its self-developed instruction set and GPU parallel computing engine, MXMACA integrates mainstream algorithm frameworks, computing libraries, toolchains, and development environments to deliver a full-stack software solution for MetaX's GPU series, covering development, debugging, and performance optimization.

MXMACA actively embraces open-source frameworks, libraries, and models, continuously optimizing for performance and scalability. As a founding member of the CCF AI Infra Working Group, MetaX collaborates with industry, academia, and developers to foster innovation through contests, training, incubation, and investment programs, advancing China's AI ecosystem.