The field of generative 3D and 4D modeling is rapidly evolving, with a focus on creating more realistic and physically plausible models. Researchers are exploring new methods for generating 3D objects and scenes from minimal input, such as 2D images or map footprints. These approaches often involve hierarchical frameworks, physics-based simulations, and advanced generative models. The goal is to produce models that are not only visually realistic but also functional and interactive. Noteworthy papers in this area include:

• In-2-4D, which generates 4D models from two single-view images.
• SOPHY, which produces simulation-ready objects with physical materials.
• HiScene, which creates hierarchical 3D scenes with isometric view generation.