Hy-3D (Hunyuan3D) is an advanced generative AI tool developed by Tencent’s Hunyuan team that creates detailed, PBR-ready 3D assets from simple text prompts or a single 2D image. Using a two-stage diffusion transformer architecture, it first constructs coherent mesh geometry and then synthesizes realistic textures (including albedo, normal, roughness, and metallic maps). The result is high-fidelity, production-ready 3D objects that are suitable for game development, 3D printing, VFX, e-commerce, and more. The tool is open-source under an Apache 2.0 license, free to use, and can run locally with modest GPU requirements (around 6 GB VRAM).