LiveSecBench is a newly developed AI safety benchmark tailored specifically for Chinese-language large language models (LLMs). Its primary purpose is to assess these models on critical dimensions such as legality, ethics, and privacy, ensuring alignment with the distinct requirements of the Chinese context. A notable feature of LiveSecBench is its dynamic update mechanism, which allows it to continuously incorporate emerging threats and challenges, thereby maintaining its relevance over time. This adaptability makes it an essential tool for developers aiming to create safe and culturally appropriate AI applications in China. The benchmark’s significance is underscored by its positive reception within the AI community, highlighting its role in advancing responsible AI deployment. Furthermore, multiple recent studies and reports have mirrored its objectives and design, reflecting a growing consensus on the need for culturally relevant safety evaluations in AI. Overall, LiveSecBench represents a vital step toward enhancing the safety and ethical standards of LLMs operating within the Chinese linguistic and cultural environment.