On the evening of July 28, Beijing Knowledge Atlas Technology Joint Stock Co., Ltd. (Z.ai) released a next-generation flagship large-scale model, GLM-4.5. Designed as a foundation model for agent-based applications, it integrates and advances general capabilities such as complex reasoning, code generation, and agent interaction. Comprehensive test results show that it is among the world's leading models. The launch marks a significant step in AI's progress toward artificial general intelligence (AGI).

[Photo via VCG]

Its release quickly caught the attention of international media. Within two hours, X (formerly Twitter) featured the model on its homepage. CNBC reported: "Chinese companies are making smarter artificial intelligence models that are increasingly cheaper to use, echoing key aspects of DeepSeek's market-shaking breakthrough." VentureBeat, a renowned US-based tech news outlet, noted: "GLM-4.5's launch gives enterprise teams a viable, high-performing foundation model. It's a compelling option for teams balancing innovation, performance, and operational constraints." Meanwhile, Techi, a US-based tech website, posted that with  an increasing number of firms selling open-source models at affordable prices, China is now starting to become a key player in the global race for Artificial Intelligence leadership.

Currently, this high-performance, cost-efficient, open-source model is publicly available on the Zhipu AI Open Platform. Its open-source version will also be released on Hugging Face and ModelScope.

 (Source: Beijing Fabu)