On August 11, ZhiPu unveiled its latest generation of visual understanding model, GLM-4.5V, and announced its open-source availability. GLM-4.5V builds upon ZhiPu's flagship text-based model, GLM-4.5-Air, continuing the technological trajectory established by GLM-4.1V-Thinking. Boasting a total of 106 billion parameters and 12 billion active parameters, GLM-4.5V achieves state-of-the-art (SOTA) performance among open-source models of its class across 41 public visual multimodal benchmarks. These benchmarks encompass a wide range of tasks, including image and video understanding, document analysis, and GUI Agent interactions. Furthermore, GLM-4.5V introduces an innovative thinking mode switch, empowering users to tailor the model's level of depth in processing according to their specific needs.