Rockchip released a record sheet of investor relations activities. The company can provide downstream customers and ecosystem partners with different computing power chips from 0.2 tops to 6 tops. Among them, the RK3588 and RK3576 have 6 topSNPU processing units, which can support mainstream 0.5B to 3B parameter level model deployment on the end side, implement functions such as translation, summary, question and answer through large language models, and can achieve multi-modal search and recognition, effectively solving the pain points of different AIoT scenarios and improving the product user experience. Recently, many models have appeared in the industry to explore the implementation of generative AI application scenarios through various smart terminal hardware. Currently, customers in various fields have developed new hardware that supports large AI models on the end side based on Rockchip microcontroller chips, such as AI learning machines, dictionary pens, smart speakers, computing power terminals, conference hosts, etc.