VGGT-World: Transforming VGGT into an Autoregressive Geometry World Model
This work addresses the challenge of efficient and accurate 3D world modeling for applications like autonomous driving and robotics, representing an incremental advance by building on existing geometry-foundation models.
The paper tackles the problem of geometrically inconsistent predictions in world models by introducing VGGT-World, which forecasts the evolution of frozen geometry-foundation-model features instead of generating video frames, achieving significant improvements in depth forecasting with 3.6-5 times faster runtime and only 0.43B trainable parameters.
World models that forecast scene evolution by generating future video frames devote the bulk of their capacity to photometric details, yet the resulting predictions often remain geometrically inconsistent. We present VGGT-World, a geometry world model that side-steps video generation entirely and instead forecasts the temporal evolution of frozen geometry-foundation-model (GFM) features. Concretely, we repurpose the latent tokens of a frozen VGGT as the world state and train a lightweight temporal flow transformer to autoregressively predict their future trajectory. Two technical challenges arise in this high-dimensional (d=1024) feature space: (i) standard velocity-prediction flow matching collapses, and (ii) autoregressive rollout suffers from compounding exposure bias. We address the first with a clean-target (z-prediction) parameterization that yields a substantially higher signal-to-noise ratio, and the second with a two-stage latent flow-forcing curriculum that progressively conditions the model on its own partially denoised rollouts. Experiments on KITTI, Cityscapes, and TartanAir demonstrate that VGGT-World significantly outperforms the strongest baselines in depth forecasting while running 3.6-5 times faster with only 0.43B trainable parameters, establishing frozen GFM features as an effective and efficient predictive state for 3D world modeling.