Zhipu AI's GLM-5.2 goes open source, stock surges 48%
Synopsis
Key Takeaways
Zhipu AI, the Beijing-based generative AI company listed in Hong Kong as Knowledge Atlas Technology, saw its shares rocket as much as 48 per cent to HK$1,620 on Monday, 15 June 2026, after it announced plans to release its most powerful large language model, GLM-5.2, as open-source software later this week. The stock pulled back to HK$1,404 by the noon break, still representing a dramatic single-session move.
The announcement
Zhipu AI confirmed on Saturday that the full release of GLM-5.2 will feature a massive 1-million-token context window — a specification that places it among the most capable open models globally. The model will be accessible through the company's new GLM Coding Plan subscription tier, designed specifically for developers and coding workflows.
The pricing edge
The GLM Coding Plan is priced at roughly a tenth of Anthropic's premium Claude Code and Claude Max tiers, according to the company. This aggressive pricing strategy positions Zhipu AI directly against Western AI incumbents, targeting cost-sensitive developers and enterprises seeking high-performance alternatives at a fraction of the cost.
Why it matters
The open-source move reflects a broader pattern among Chinese AI developers: leveraging openness and competitive pricing to capture users priced out of premium Western platforms. By releasing GLM-5.2 openly, Zhipu AI lowers the barrier for global adoption and accelerates community-driven development, potentially expanding its developer ecosystem far beyond its domestic base. The strategy mirrors moves by rivals such as DeepSeek, which similarly used open-source releases to gain rapid international traction.
Market reaction
Knowledge Atlas Technology's stock has now climbed more than 780 per cent since its blockbuster initial public offering in early January 2026, making it one of the standout performers on the Hong Kong exchange this year. Monday's intraday surge underscores how sensitive AI-sector valuations remain to model capability announcements, particularly those that shift the open-source competitive landscape.
What's next
The full public release of GLM-5.2 is expected later this week, with developer uptake and benchmark comparisons against leading Western models set to be closely watched. How Anthropic and other premium AI providers respond — whether through pricing adjustments or accelerated model releases — will be a key indicator of how intensely the open-source price war is beginning to bite.