Tencent Open-Sources First 3D World Generation AI Model

Tencent’s Hunyuan AI lab has released Hunyuan3D World Model 1.0, billed as the industry’s first open-source model that can turn a text or image prompt into a fully explorable three-dimensional environment. The code went live on Sunday after the company announced the project at the World Artificial Intelligence Conference 2025 in Shanghai. Built on the lab’s new Hunyuan 3D v2.5 sparse, 3D-native architecture, the model outputs scenes that remain editable within standard computer-graphics pipelines, allowing developers to fine-tune assets for videogames, virtual-reality applications and simulation software. By releasing the technology under an open-source licence, Tencent is seeking to broaden developer adoption and strengthen its position in China’s fast-growing generative-AI market.