cs.CVNov 19, 2026

SpecWM: Turning Verification into Refinement for Streaming World Models

Authors: Jiaming Li, Lingzhen Li, Xiaze Yang, Chenyu Zhu, Kaiqi Liu, Xiaotong Zhao, Songlian Li, Toyota Li, 

Organizations: Huazhong University of Science and Technology · Tencent · Peking University · The Hong Kong University of Science and Technology

Abstract

Recently, video world models have advanced interactive generation, with large models offering higher quality and smaller models enabling faster inference. Existing speculative methods combine their strengths through fast draft generation and selective target regeneration guided by rewards on decoded frames. However, verification covers only the external scorer's evaluation dimensions and adds decoding and scoring costs. Moreover, these scores provide no correction, requiring costly target regeneration for rejected drafts. To address these limitations, we propose Speculative World Model (SpecWM), a unified, training-free inference framework for streaming world models. It verifies draft chunks using the target's latent velocity field. Verification also yields a reusable correction; two thresholds select Accept, Refine, or Reject. Specifically, one-step refinement reuses it to improve perceptual quality, while three-step refinement strengthens camera following and trajectory consistency. These findings motivate transition-aware refinement. On WBench Navigation, SpecWM improves visual quality at 1.63×\times speedup over the target and raises NavScore from the draft's 75.99 to 85.63, achieving the best quality-speed-control trade-off. Additionally, our LongSpecBench shows sustained quality and control over long, multi-turn rollouts, while experiments on Hy-Nomas validate SpecWM across backbones and denoising budgets.

Figures & tables

Explore similar work

CardsList
  1. World in World: Explore the World with World Models

    Sep 10, 2026Chenxi Song, Yanming Yang, Chi ZhangCamera-Controlled Video GenerationLong-Horizon Video Generation

  2. Precise Editing and Flexible Referencing for Interactable Worlds

    Sep 28, 2026Xinyao Liao, Xianfang Zeng, Zhu Liang +5Instruction-Based Video EditingAutoregressive Video Generation