# Tencent-Hunyuan/HunyuanWorld-Mirror

[ICML 2026] WorldMirror: Fast and Universal 3D reconstruction model for versatile tasks

Repository: https://github.com/Tencent-Hunyuan/HunyuanWorld-Mirror
Canonical: https://ross.abutalabs.com/products/hunyuanworld-mirror
Homepage: https://3d-models.hunyuan.tencent.com/world/
Language: Python
License: NOASSERTION
License Family: other
Topics: 3d, 3d-reconstruction, aigc, hunyuan, hunyuan3d, image-to-3d, scene-generation, world-model, world-models
Last push: 2026-05-27T03:22:57+00:00

## Health v2 (maintenance only)
Score: 54/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 84, release rhythm 35, longevity 22
- inputs: {"age_days": 321, "days_push": 98, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1191, forks 120 (observed 2026-08-28T04:03:56.139740+00:00)

## What it is
HunyuanWorld-Mirror is a feed-forward 3D reconstruction model from Tencent that predicts camera poses, intrinsics, depth maps, point clouds, surface normals, and 3D Gaussians in a single pass. It supports versatile 3D tasks including image-to-3D and scene generation, with inference, training, and evaluation code plus pretrained weights.

## Use cases
- reconstruct 3d scenes from images
- estimate camera poses and depth from multi-view photos
- generate 3d gaussians from a single image
- convert images to point clouds
- predict surface normals for 3d scenes
- build a world model for scene generation

## When to choose
- you need fast feed-forward 3D reconstruction without per-scene optimization
- you want multiple 3D outputs (depth, normals, point clouds, Gaussians) from one model
- you need camera parameter estimation from uncalibrated images

## When to avoid
- you need classical photogrammetry with precise metric accuracy
- you lack a GPU for inference
- you need a permissively licensed model for commercial use without checking the custom license

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, computer-vision, image-processing, graphics, simulation
- domain: computer-vision, graphics, artificial-intelligence, deep-learning, simulation
- platform: python, cross-platform
- tags: 3d-reconstruction, world-model, point-cloud, depth-estimation, camera-pose-estimation, 3d-gaussians, feed-forward-model, image-to-3d, scene-generation, hunyuan, gpu, linux

## Member repositories
- Tencent-Hunyuan/HunyuanWorld-Mirror (main) score 54

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:56.139740+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:22:41.268525+00:00, confidence not recorded.
  - readme: https://github.com/Tencent-Hunyuan/HunyuanWorld-Mirror (fetched 2026-08-28T04:03:56.139740+00:00, sha aa3449231a04)
  - homepage: https://3d-models.hunyuan.tencent.com/world/ (fetched 2026-08-29T12:30:10.550010+00:00, sha 64d8acac6828)
- Data as of 2026-08-30T08:39:29.467469+00:00.