Multimodal Large Language Models (MLLMs) have excelled in 2D image-text comprehension and image generation. Still, their understanding of the 3D world needs to be improved, limiting progress in 3D ...