Zidong Taichu Open-Sources ZDTaichu5.0-9B: Leading in Spatial Embodied AI Without Compromising General Capabilities

Zidong Taichu has open-sourced ZDTaichu5.0-9B, a general multimodal large model that ranks first in eight out of nine international spatial understanding benchmarks, demonstrating a significant lead in spatial embodied capabilities. The model also maintains top-tier performance in general tasks such as image-text understanding, OCR, mathematical reasoning, and coding, achieving breakthroughs in both spatial and general abilities.

+13