VGGT-World: Transforming VGGT into an Autoregressive Geometry World Model

arXiv 2026 46.3 method

TLDR

VGGT-World repurposes frozen VGGT features as autoregressive latent state to forecast scene geometry without video generation, achieving efficient depth forecasting on KITTI, Cityscapes, and TartanAir.

Reasoning

The paper introduces a novel geometry world model that avoids video generation by predicting latent GFM token trajectories, with technical contributions addressing flow matching collapse and exposure bias. Strengths include strong empirical results and efficiency, but the abstract focuses mainly on depth forecasting and lacks evidence of broader interactive or generative capabilities.

Read-first score

Read-first score 46.3, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 26.

Recency 8%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Methodology quality 25%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=baseline,experiment,metric,result

Topical relevance 42%

37.1

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 609.

Keyword Scores

world model

world dynamics prediction

generative world model

world simulator

interactive world model

video world model

model-based reinforcement learning world model

Resources

Paper