Zidong Taichu Open-Sources ZDTaichu5.0-9B, a 9B-Parameter Multimodal Model

On September 15, Zidong Taichu officially open-sourced ZDTaichu5.0-9B, a general multimodal large model designed for the physical world. The model, which has approximately 9 billion parameters, supports single-image, multi-image, long-sequence video and arbitrary-resolution visual inputs. According to the team, it focuses on addressing challenges in real-world physical scenarios, including spatial perception, cross-view transformation and embodied task planning.