WorldMirror: Universal 3D World Reconstruction with Any-Prior Prompting

We present WorldMirror, a unified feed-forward model for comprehensive 3D geometric prediction tasks. Unlike existing methods constrained to image-only inputs or customized for a specific task, our framework flexibly integrates diverse geometric priors, including camera poses, intrinsics, and depth maps, while simultaneously generating multiple 3D representations: dense point clouds, multi-view depth maps, camera parameters, surface normals, and 3D Gaussians. Remarkably, prior injection yields universal gains across all tasks, suggesting that input flexibility and multi-task prediction are mutually reinforcing. WorldMirror achieves state-of-the-art performance across diverse benchmarks from camera, point map, depth, and surface normal estimation to novel view synthesis, while maintaining the efficiency of feed-forward inference. Code and model weights are publicly available at https://github.com/Tencent-Hunyuan/HunyuanWorld-Mirror.

WorldMirror is a fast feed-forward model that predicts multiple 3D outputs at once, including point clouds, multi-view depth, cameras, normals, and 3D Gaussians, instead of using separate tools for each task. It can take optional inputs such as poses, intrinsics, and depth, and these priors broadly improve performance. It reports strong results on geometry and novel-view benchmarks while remaining efficient at inference. Code and weights are available on GitHub (HunyuanWorld-Mirror).