Wonder: Video World Model Done Better
2026-07-28 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionGraphics
AI summaryⓘ
The authors introduce Wonder, a video model that creates interactive, playable worlds from images or videos, letting users move the camera and explore scenes in real time. They developed a new way to link camera movement with visual cues for better control, and an efficient memory system to handle long video sequences without slowing down. Their training methods help the model follow camera commands while keeping video quality and diversity over time. This approach works not only for generating videos from images but also for modifying existing videos interactively.
video world modelcamera controldense coordinate fieldsparse attentionmemory mechanismself-forcing distillationvideo synthesisimage-to-video generationreal-time renderinginteractive exploration
Authors
Jiacong Xu, Hanwen Jiang, Zhixin Shu, Kalyan Sunkavalli, Vishal M. Patel, Yiqun Mei
Abstract
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.