# nv-tlabs/GEN3C

[CVPR 2025 Highlight] GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control

Repository: https://github.com/nv-tlabs/GEN3C
Canonical: https://ross.abutalabs.com/products/gen3c
Homepage: https://research.nvidia.com/labs/toronto-ai/GEN3C/
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Topics: 3d-graphics, camera-control, video-diffusion-model
Last push: 2026-06-15T19:11:34+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 87, release rhythm 35, longevity 39
- inputs: {"age_days": 548, "days_push": 79, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1409, forks 80 (observed 2026-08-28T04:04:38.665892+00:00)

## What it is
GEN3C is NVIDIA's research codebase for a generative video model that achieves precise camera control and temporal 3D consistency using a 3D point-cloud cache. It supports video generation from single images, sparse-view images, or dynamic videos with user-specified camera trajectories.

## Use cases
- generate video with precise camera control from a single image
- novel view synthesis from sparse-view images
- extend monocular dynamic video with new camera trajectories
- create temporally consistent video without objects popping in and out
- render dolly zoom and cinematic camera effects on generated video
- video-to-video generation with multi-view inference

## When to choose
- you need 3D-consistent video generation with exact camera pose control
- you want state-of-the-art sparse-view novel view synthesis
- you are doing research on world-consistent video diffusion models

## When to avoid
- you need a production-ready turnkey video generation product
- you lack a GPU or cannot run diffusion model inference
- you need simple text-to-video without camera control

## Facets
- artifact type: library
- maturity: active
- function: video-processing, machine-learning, deep-learning, computer-vision, graphics
- domain: computer-vision, deep-learning, graphics, artificial-intelligence
- platform: python
- tags: video-diffusion, camera-control, novel-view-synthesis, 3d-cache, world-model, research-code, cvpr-2025, gpu, linux

## Member repositories
- nv-tlabs/GEN3C (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:38.665892+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 2026-08-30T04:38:33.212525+00:00, confidence not recorded.
  - readme: https://github.com/nv-tlabs/GEN3C (fetched 2026-08-28T04:04:38.665892+00:00, sha fcc91891bdc1)
  - homepage: https://research.nvidia.com/labs/toronto-ai/GEN3C/ (fetched 2026-08-29T11:52:05.988189+00:00, sha 4f16dba6e4ff)
- Data as of 2026-08-30T08:39:29.467469+00:00.