This content describes how an AI agent successfully constructed a 3D Paris gallery. The achievement was made by chaining two Hugging Face Spaces, illustrating the capabilities of agents in complex generative tasks.
Tencent has released HY-World 2.0, the first open-source state-of-the-art 3D world model that generates real, persistent, and editable 3D assets. Unlike video-only models, it offers true 3D consistency, native physics, and is fully importable into major game engines and 3D software.
Microsoft's TRELLIS.2 is a 4B-parameter open-source model for high-fidelity image-to-3D generation, capable of producing up to 1536³ PBR textured assets using native 3D VAES and a novel O-Voxel structure. It offers an efficient, scalable solution for generating detailed 3D assets with full PBR materials.
Um desenvolvedor criou um aplicativo de IA 'waifu' para ensino de idiomas, utilizando Gemma-4, Omnivoice TTS e modelagem 3D. O app, com recursos como chamadas de voz/vídeo, impressionou o criador pela capacidade de Gemma-4 de seguir prompts sem censura.
MashuPack is a web-based, no-code tool designed for creating custom 3D models and animations by importing, modifying, and combining assets. Technically, it utilizes modern web frameworks, a RESTful API, and WebGL for in-browser rendering.
AI is revolutionizing 3D model generation, drastically reducing creation time from days to minutes using Text-to-3D or Image-to-3D techniques. This innovation democratizes 3D design, making it accessible for hobbyists and professionals for custom creations and rapid prototyping.