Shanghai Wuwen Xinqiong Intelligent Technology Co., Ltd. and Shanghai Xiyu Technology (MiniMax) have recently signed a strategic cooperation agreement. Both AI companies based in Shanghai have expanded their collaboration in four directions: co-optimizing the model inference efficiency system, jointly building large model Token services and AI infrastructure, joint innovation in application paradigms, and large-scale expansion of intelligent scenarios in industries.

Translate this cooperation framework: Wuwen Xinqiong excels in computing power scheduling and inference infrastructure, while MiniMax holds self-developed large models and massive application traffic. The former helps the latter run models more efficiently and cost-effectively, while the latter provides the former with real business scenarios and large volumes of Token consumption. The two companies are integrating from the bottom layer of inference efficiency, to the middle layer of Token services and infrastructure, and up to the upper layer of industry scenarios, effectively linking two key players in the "training-inference-deployment" chain of domestic large models. In the current context where both computing costs and application deployment face pressure, such upstream and downstream collaboration has become an increasingly common survival strategy within the domestic large model ecosystem.