智谱：视觉推理模型GLM-4.5V正式上线并开源

Core Viewpoint - The company has launched the open-source visual reasoning model GLM-4.5V, which features a total of 106 billion parameters and 12 billion active parameters, and is available on both the Modao community and Hugging Face [1] Group 1: Product Details - GLM-4.5V is based on the company's next-generation flagship text foundation model GLM-4.5-Air [1] - The model continues the technical route of GLM-4.1V-Thinking [1] - It achieves state-of-the-art (SOTA) performance among open-source models in 41 public visual multimodal benchmarks, covering tasks such as image, video, document understanding, and GUI Agent [1]