Oppo announces on-device AI breakthrough with Mixture of Experts architecture

Oppo has announced it is the first company to implement Mixture of Experts (MoE) AI architecture on devices. This technology improves processing efficiency by activating specialized sub-models for specific tasks, enhancing performance while reducing energy consumption. Laboratory tests show that on-device MoE can accelerate AI tasks by about 40%, leading to faster responses and longer battery life. It also enhances privacy by processing more tasks locally, rather than relying on cloud computing. Oppo collaborated with chipset providers to address the challenges of large AI models on devices with limited resources. The company has filed over 5,860 patent applications in AI technology, highlighting its commitment to advancing on-device AI.