A Frame is Worth One Token: Efficient Generative World Modeling with Delta Tokens
2026-04-06 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors address the problem of predicting many possible future video frames efficiently. They create DeltaTok, a tool that summarizes changes between video frames into simple tokens, and DeltaWorld, a model that uses these tokens to generate multiple possible futures quickly. This approach compresses video data drastically, enabling the generation of diverse future predictions using far fewer resources than previous methods. Their experiments show that DeltaWorld predicts future video frames more accurately and efficiently than existing models.
video world modelingvision foundation model (VFM)tokenizergenerative modelmultihypothesis predictionfeature differencespatio-temporal representationFLOPsdense forecasting
Authors
Tommie Kerssies, Gabriele Berton, Ju He, Qihang Yu, Wufei Ma, Daan de Geus, Gijs Dubbelman, Liang-Chieh Chen
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. Code and weights: https://deltatok.github.io.