Zhipu AI released GLM-4.5V, best 100B visual model

Zhipu AI released GLM-4.5V, an open-source visual reasoning model, claiming it's the best performing 100B-class model globally. The model, with 106 billion total parameters, is available on Hugging Face and Modao Community. It excels in image, video, and document understanding, achieving state-of-the-art results on 41 benchmarks. GLM-4.5V features a "thinking mode" switch for flexible response times and includes a desktop assistant application for developers to build multimodal applications.