Shanghai (Gasgoo)- On September 19, Huixi Technology Co.,Ltd. ("HUIXI") and ModelBest, both of which are Chinese companies, formally signed a strategic cooperation agreement. By combining HUIXI's strengths in high-performance AI chips with ModelBest's expertise in large-model algorithms, the partners plan to advance innovation in edge AI and accelerate the deployment of powerful AI systems across a wide range of devices.
ModelBest has established itself as a leader in edge AI, focusing on the research and commercialization of large-scale models designed for on-device use. Its MiniCPM series — nicknamed "Little Cannon" — is known for high knowledge density and fast response, achieving strong benchmark results. The series are already deployed in cars, smartphones, PCs, and embodied intelligence systems, supporting a wide array of chips and platforms with efficient, adaptable performance.
HUIXI, meanwhile, is developing high-compute AI chips and corresponding hardware platforms, paired with accessible and open software toolchains. The company positions its products as a cost-efficient, high-performance foundation for embodied intelligence and autonomous driving. Its chips stand out for computational efficiency, energy control, and cost optimization, making them well-suited to edge computing scenarios requiring real-time and highly reliable processing.
Under the agreement, the two companies will collaborate across a broad spectrum of edge AI use cases. Their joint efforts include optimizing synergy between large models and AI chips, conducting co-debugging, deployment validation, and co-development. This "algorithm-chip" integration is intended to improve performance and ensure their technologies are finely tuned for commercialization in diverse scenarios.
Industry observers see 2025 as the starting point for large-scale adoption of on-device intelligence, as technology advances, use cases expand, and policy support grows. The partnership between ModelBest and HUIXI represents a convergence of leading players in algorithms and high-performance chips. Both companies intend to form a long-term alliance, pooling resources and complementary strengths to build a more complete edge AI ecosystem — one aimed at wider, safer, and more efficient adoption across industries.
