Metraipolis › archive › dc_c70f9ae7ddfb

WorldProbe: When Video World Models Fall Apart under Long-Horizon Spatial Exploration — arXiv:2610.39721

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.39721
License (as archived)
CC0-1.0
Archived
2026-11-01T02:42:09.043Z
SHA-256 of the archived text
b41c00bb70f916ca74e0a27cb588fdb6e0fce0546074e49bf15d96a9ae6057f0
Stable URL of this page
https://metraipolis.com/documento/dc_c70f9ae7ddfb

How to cite

WorldProbe: When Video World Models Fall Apart under Long-Horizon Spatial Exploration — arXiv:2610.39721 — arXiv (en), https://arxiv.org/abs/2610.39721 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c70f9ae7ddfb (Metraipolis archive dc_c70f9ae7ddfb, sha256 b41c00bb70f916ca74e0a27cb588fdb6e0fce0546074e49bf15d96a9ae6057f0)
BibTeX
@misc{dcc70f9ae7ddfb,
  title = {WorldProbe: When Video World Models Fall Apart under Long-Horizon Spatial Exploration — arXiv:2610.39721},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.39721},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c70f9ae7ddfb; sha256 b41c00bb70f916ca74e0a27cb588fdb6e0fce0546074e49bf15d96a9ae6057f0},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39721. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Video world models generate future visual observations from camera or action inputs, with the goal of synthesizing persistent 3D worlds that can be explored across time and space. Existing benchmarks typically evaluate short interactions or rely on camera trajectories fixed before generation. The performance of video world models under long-horizon exploration with extensive camera motion remains under-explored. We introduce WorldProbe, a diagnostic benchmark that uses a goal-conditioned multimodal agent to adapt camera actions online during generation. WorldProbe comprises 164 test cases across four probing families. Across 10 state-of-the-art models, we generate more than 1,600 rollouts totaling approximately 800 minutes, with horizons up to 180 seconds. Our evaluation across four dimensions reveals both the degradation of local measures and the emergence of long-range inconsistencies. We find that recent models sustain visual and content quality for longer, while long-range memory and evolution remain weak. These results expose a gap between locally stable generation and persistent world consistency that is largely missed by existing benchmarks. Code and benchmark resources are available at this https URL .

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.39721 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-01T02:16:43.950Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).