The model will understand text, image, video, or 3D asset prompts, meaning users can drop in an existing 3D game asset and ask the AI to make something similar, change it, or expand upon it, for ...