Wonder: Video World Model Done Better

Hugging Face Daily Papers Papers
video-world-model real-time camera-control sparse-attention memory-mechanism distillation world-exploration

Summary

Wonder is a general-purpose video world model that enables real-time, camera-controllable world exploration from an image or conditional video. It introduces camera conditioning via dense coordinate fields, a sparse attention memory mechanism, and techniques to improve distillation, allowing minute-scale video generation at 16 FPS.

We present Wonder, a general-purpose video world model for real-time, camera-controllable world exploration. Given an image or a conditional video, Wonder constructs a playable world where users can navigate interactively by moving the camera, discovering unseen regions, and revisiting previously observed areas in real time and over a long-term horizon. Achieving this capability requires a system-level co-design of control method, memory mechanism, and training strategy. We introduce a novel camera conditioning with a dense coordinate field whose renderings provide spatially aligned motion and orientation cues, allowing the model to interpret camera motion directly as visual evidence. To support fast and precise memory retrieval over a growing generation context, we propose an efficient sparse attention-based memory mechanism, enabling the model to selectively attend to a small set of relevant context tokens at inference time, regardless of actual context length. We further develop several techniques to rectify the self-forcing-style distillation pipeline, improving the student model's ability to respect control signals, as well as maintaining diverse generation modes and long-term memory from the teacher. Together, these components enable Wonder to synthesize diverse, minute-scale videos at 16 FPS while preserving coherent geometry, appearance, and dynamics across long rollouts. Beyond image-to-video generation, Wonder naturally supports video-conditioned generation, allowing existing dynamic scenes to be re-shot in real time.
Original Article
View Cached Full Text

Cached at: 07/29/26, 03:50 AM

Paper page - Wonder: Video World Model Done Better

Source: https://huggingface.co/papers/2607.26037

Abstract

WepresentWonder,ageneral-purposevideoworldmodelforreal-time,camera-controllableworldexploration.Givenanimageoraconditionalvideo,Wonderconstructsaplayableworldwhereuserscannavigateinteractivelybymovingthecamera,discoveringunseenregions,andrevisitingpreviouslyobservedareasinrealtimeandoveralong-termhorizon.Achievingthiscapabilityrequiresasystem-levelco-designofcontrolmethod,memorymechanism,andtrainingstrategy.Weintroduceanovelcameraconditioningwithadensecoordinatefieldwhoserenderingsprovidespatiallyalignedmotionandorientationcues,allowingthemodeltointerpretcameramotiondirectlyasvisualevidence.Tosupportfastandprecisememoryretrievaloveragrowinggenerationcontext,weproposeanefficientsparseattention-basedmemorymechanism,enablingthemodeltoselectivelyattendtoasmallsetofrelevantcontexttokensatinferencetime,regardlessofactualcontextlength.Wefurtherdevelopseveraltechniquestorectifytheself-forcing-styledistillationpipeline,improvingthestudentmodel’sabilitytorespectcontrolsignals,aswellasmaintainingdiversegenerationmodesandlong-termmemoryfromtheteacher.Together,thesecomponentsenableWondertosynthesizediverse,minute-scalevideosat16FPSwhilepreservingcoherentgeometry,appearance,anddynamicsacrosslongrollouts.Beyondimage-to-videogeneration,Wondernaturallysupportsvideo-conditionedgeneration,allowingexistingdynamicscenestobere-shotinrealtime.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2607\.26037

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2607.26037 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.26037 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.26037 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Holo-World: Unified Camera, Object and Weather Control for Video World Model

Hugging Face Daily Papers

Holo-World presents a unified controllable video world model that generates videos from a single image with explicit control over camera, object motion, and weather. It introduces a novel dataset and techniques to preserve scene structure while transferring to target weather states.

Qwen-RobotWorld Technical Report: Unifying Embodied World Modeling through Language-Conditioned Video Generation

Hugging Face Daily Papers

Qwen-RobotWorld is a language-conditioned video world model that predicts future visual trajectories across multiple robotic domains using a double-stream diffusion transformer and an 8.6M video-text corpus. It unifies embodied world modeling for robotic manipulation, autonomous driving, indoor navigation, and human-to-robot transfer, achieving top benchmarks on EWMBench and DreamGen Bench.