Zhipu AI, the Chinese artificial intelligence developer also known as Knowledge Atlas, saw its Hong Kong-listed shares move sharply higher after unveiling a new large language model tailored to agent-style AI workflows.
The company introduced GLM-5-Turbo, described on its developer platform as a foundation model explicitly trained for OpenClaw scenarios. According to the company’s platform notes, the model is intended to strengthen capabilities such as tool invocation, following instructions and executing long-chain tasks that span multiple steps.
Market reaction to the launch was immediate: shares of the firm rose as much as 16% to HK$615 as of 04:11 GMT.
OpenClaw is identified by Zhipu as a framework for AI assistants that can operate locally or across enterprise systems while interacting with a range of tools and messaging platforms. GLM-5-Turbo supports several features aimed at agent deployment, including real-time streaming responses, structured output formats and integration with external toolsets and data sources, which the company says should make embedding the model into enterprise environments more straightforward.
The release underscores Zhipu’s strategic emphasis on building AI infrastructure for autonomous agents. The company framed the launch as part of a broader move into agent-focused tooling, an area the article noted is widely regarded within the technology sector as the next phase of generative AI development.
Details published on the developer platform emphasize that GLM-5-Turbo’s architecture and training are oriented around scenarios where models must coordinate with external tools and manage multistage processes - capabilities central to agent-style applications. The platform material also highlights the model’s output and connectivity features intended to facilitate enterprise integration.
While the announcement drew investor interest and a sizable intraday jump in the stock, the information provided by the company focuses on technical capabilities and intended use cases rather than commercial terms or deployment timelines.
Context note: The company provided the technical and product descriptions via its developer platform; market pricing cited reflects the share movement reported following the product launch.