Jefferies has released a research report stating that the GLM 5.2 model from KNOWLEDGE ATLAS (02513) has achieved a global ranking of third place on the Artificial Analysis Model Intelligence Index. This marks the first time a Chinese model has broken into the top three positions, with specific rankings of fourth globally in programming capabilities and second in agent intelligence, representing a significant development milestone for China's AI sector.

However, Jefferies also cautioned that U.S. access restrictions on Anthropic's Fable 5 model may be difficult to sustain over the long term, and developers have a variety of alternative options and open-source models available for selection. Currently, KNOWLEDGE ATLAS is considered overvalued; based on its guidance to achieve $1 billion in annualized recurring revenue (ARR) by the end of 2026, its current price-to-ARR multiple is as high as 94 times, compared to a multiple of only about 18 times for Anthropic.