DreamZero: 14B Autoregressive Video Diffusion Model Enables 7Hz Real-Time Closed-Loop Robot Control
du_yilun · x · 2026-08-13
The author shares a new paper, World Action Models are Zero-shot Policies, and its code, introducing DreamZero, a World Action Model (WAM) built upon a pretrained video diffusion backbone.
- Core Strength: By jointly modeling video and action to learn physical dynamics without relying on repetitive demonstrations, DreamZero achieves over a 2x improvement in generalization to new tasks and environments compared to state-of-the-art Vision-Language-Action (VLA) models.
- System Optimization: Through model and system optimizations, the researchers enabled a 14B autoregressive video diffusion model to perform real-time closed-loop control at 7Hz.
- Cross-embodiment Transfer: It demonstrates robust cross-embodiment capabilities. Using just 10-20 minutes of video-only demonstrations from other robots or humans yields a relative improvement of over 42% on unseen tasks.
More from Embodied
- AutoAgent0: VLM-powered agentic runtime lifts long-tail driving success by 28.3pp — yining_hong · 2026-10-16
- CarryOn: dual-loop harness adds personalization and safe recovery to frozen VLA robots — Zixuan Wang · 2026-10-16
- FoveaVLA replaces wrist cameras with learned foveation, hitting 70.8% success — Haoyu_Xiong_ · 2026-10-16
- Musk teases two-way Grok x Optimus integration: Grok in the bot, bot in Grok — beffjezos · 2026-10-16
- SimVTA uses simulator-rendered video to post-train driving world action models with RL — OpenDriveLab · 2026-10-16
- Robotics data startups compared: Scale AI tops $1B revenue as 9 LLM vs robotics data suppliers benchmarked — FinanceYF5 · 2026-10-16