CVPR 2026

A Frame is Worth One Token:
Efficient Generative World Modeling
with Delta Tokens

Tommie Kerssies1,2,*, Gabriele Berton1,*, Ju He1, Qihang Yu1, Wufei Ma1,3,*
Daan de Geus2,**, Gijs Dubbelman2,**, Liang-Chieh Chen1,*
1Amazon   2Eindhoven University of Technology   3Johns Hopkins University
*Work done while at Amazon. **Equal advising.
Paper Code Models
Outline of DeltaWorld
Figure 1. Existing generative world models require many forward passes and many spatial tokens per frame. DeltaWorld generates diverse futures in a single forward pass with one delta token per frame.

Abstract

Anticipating diverse future states is a central challenge in video world modeling. Discriminative world models produce a deterministic prediction that implicitly averages over possible futures, while existing generative world models remain computationally expensive. Recent work demonstrates that predicting the future in the feature space of a vision foundation model (VFM), rather than a latent space optimized for pixel reconstruction, requires significantly fewer world model parameters. However, most such approaches remain discriminative. In this work, we introduce DeltaTok, a tokenizer that encodes the VFM feature difference between consecutive frames into a single continuous "delta" token, and DeltaWorld, a generative world model operating on these tokens to efficiently generate diverse plausible futures. Delta tokens reduce video from a three-dimensional spatio-temporal representation to a one-dimensional temporal sequence, for example yielding a 1,024x token reduction with 512x512 frames. This compact representation enables tractable multi-hypothesis training, where many futures are generated in parallel and only the best is supervised. At inference, this leads to diverse predictions in a single forward pass. Experiments on dense forecasting tasks demonstrate that DeltaWorld forecasts futures that more closely align with real-world outcomes, while having over 35x fewer parameters and using 2,000x fewer FLOPs than existing generative world models.

Performance comparison
Figure 2. Compared to Cosmos, DeltaWorld forecasts futures that better align with real-world outcomes while having over 35x fewer parameters and using 2,000x fewer FLOPs.

Citation

@inproceedings{kerssies2026deltatok,
  title     = {A Frame is Worth One Token: Efficient Generative World Modeling with Delta Tokens},
  author    = {Kerssies, Tommie and Berton, Gabriele and He, Ju and Yu, Qihang and Ma, Wufei and de Geus, Daan and Dubbelman, Gijs and Chen, Liang-Chieh},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2026}
}