# Tencent-Hunyuan/HY-World-2.0

HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worlds

Repository: https://github.com/Tencent-Hunyuan/HY-World-2.0
Canonical: https://ross.abutalabs.com/products/hy-world-20
Homepage: https://3d-models.hunyuan.tencent.com/world/
Language: Python
License: NOASSERTION
License Family: other
Topics: 3d, ai, worldmodel
Last push: 2026-08-12T16:15:12+00:00

## Health v2 (maintenance only)
Score: 58/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 97, release rhythm 35, longevity 10
- inputs: {"age_days": 145, "days_push": 21, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2571, forks 214 (observed 2026-08-28T04:07:01.456264+00:00)

## What it is
HY-World 2.0 is Tencent Hunyuan's open-source multi-modal world model framework that reconstructs, generates, and simulates 3D worlds from text, single-view images, multi-view images, or video, outputting meshes or Gaussian Splattings. The repository ships inference code and pretrained model weights (including WorldMirror 2.0, HY-Pano 2.0, and WorldStereo 2.0) for a staged pipeline covering panorama generation, world reconstruction, and stereo synthesis.

## Use cases
- generate navigable 3D scenes from a text prompt
- reconstruct an explorable 3D world from a single photo
- build 3D environments from multi-view images or video
- produce Gaussian Splatting or mesh representations of real scenes
- generate panoramas and stereo pairs for 3D scene pipelines
- create 3D world assets for games or simulations
- run state-of-the-art world model inference locally

## When to choose
- you need high-fidelity 3D world generation or reconstruction from text, images, or video
- you want to use pretrained Hunyuan world model weights with ready inference code
- you need mesh or Gaussian Splatting outputs for downstream 3D pipelines
- you are researching world models, 3D generation, or novel view synthesis

## When to avoid
- you need a commercial-friendly license — the repository uses a custom license (NOASSERTION) that may restrict use
- you need full training code — primarily inference code and weights are open-sourced
- you lack a capable GPU environment for large generative model inference
- you need lightweight real-time 3D editing tools rather than generation/reconstruction models

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, computer-vision, image-processing, graphics, simulation, video-processing
- domain: artificial-intelligence, deep-learning, computer-vision, image-processing, graphics, simulation
- platform: python
- tags: 3d-generation, world-model, 3d-reconstruction, gaussian-splatting, mesh-generation, text-to-3d, image-to-3d, panorama-generation, stereo-generation, scene-reconstruction, pretrained-models, inference-code, tencent-hunyuan, game-development, linux, gpu

## Member repositories
- Tencent-Hunyuan/HY-World-2.0 (main) score 58

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:01.456264+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 2026-08-30T02:23:16.788624+00:00, confidence not recorded.
  - readme: https://github.com/Tencent-Hunyuan/HY-World-2.0 (fetched 2026-08-28T04:07:01.456264+00:00, sha 8faae1735bfb)
  - homepage: https://3d-models.hunyuan.tencent.com/world/ (fetched 2026-08-29T10:05:34.938542+00:00, sha 64d8acac6828)
- Data as of 2026-08-30T08:39:29.467469+00:00.