HK Stock Market Move | KNOWLEDGE ATLAS (02513) surged over 24%, will launch a new generation model GLM-5. Institutions have high expectations for its core competitiveness.

KNOWLEDGE ATLAS (02513) rose by more than 24% on the day after its listing, reaching a high of 165 Hong Kong dollars, an increase of over 40% compared to the IPO price. As of the time of writing, it has risen by 17.34%, reaching 154.3 Hong Kong dollars, with a trading volume of 6.32 billion Hong Kong dollars. On the news front, at the beginning of 2026, the Chinese AI mega model track reached a milestone event, with KNOWLEDGE ATLAS listing on the Hong Kong Stock Exchange yesterday and becoming the "first listed company of global mega models." MiniMax, also listing today, surged by about 80% at one point. Bo Hai Securities previously pointed out that the top two startups, KNOWLEDGE ATLAS and MiniMax, racing for a Hong Kong IPO, are expected to provide key anchor points for valuation pricing and financing system establishment in the industry. Soochow stated that they are optimistic about KNOWLEDGE ATLAS's advantages in domestic mega model technology strength, open-source ecosystem layout, and the ability to localize deployment, and believe that the company is expected to benefit from the long-term trend of China's mega model industry transitioning from localized deployment to cloud services. Public information shows that KNOWLEDGE ATLAS is one of the earliest companies in China to engage in mega model research and development. According to a Frost & Sullivan report based on 2024 revenue, KNOWLEDGE ATLAS ranks first among independent general mega model developers in China and second among all general mega model developers. KNOWLEDGE ATLAS founder and chief scientist Tang Jie recently announced the upcoming launch of a new generation model GLM-5 in an internal memo. The memo also introduced the three technical directions that KNOWLEDGE ATLAS will focus on in 2026, including a new model architecture design, a more general RL (reinforcement learning) paradigm, and the exploration of continuous learning and autonomous evolution of models. These efforts are all centered on improving the fundamental model capabilities.