cs.RONov 3, 2026

PrimeSim: A Generalist World Model Simulator from Large-Scale Cross-Embodiment Demonstrations

Authors: Travis Davies, Wentao Ma, Lihao Guo, Tengfei Ma, Xiaowen Chang, Jiafeng Xu, Shilong Wang, Yiqi Deng, 

Organizations: PrimeBot Research · Southeast University · Shanghai Jiao Tong University · The Chinese University of Hong Kong · 

Abstract

A robot world model that serves as a simulator must do more than generate plausible videos: it must faithfully follow commanded actions, maintain consistency across camera views, and sustain coherent long-horizon rollouts. We present PrimeSim, an action-conditioned, multiview world model simulator that addresses these challenges through progressive midtraining and posttraining. Starting from a pretrained video generator, we midtrain on 30,110 hours of robot demonstrations spanning 22 platforms. To improve action following, we introduce a unified cross-embodiment action representation and frame-aligned dual-path conditioning to strengthen action-frame correspondence. To maintain multiview consistency, we develop view-aware interleaved attention with explicit camera identities, enabling cross-view information exchange while preserving view-specific representations. We further introduce a fast-video, slow-memory posttraining framework that transforms the bidirectional video generator into an efficient autoregressive simulator. By combining fast causal generation, self-rollout training, and hierarchical temporal memory, PrimeSim mitigates error accumulation from self-generated histories and supports coherent long-horizon simulation. PrimeSim runs at 10.1 FPS, remains stable over long-horizon rollouts on rigid and deformable manipulation tasks, and transfers to held-out platforms within 100 to 200 gradient steps where the pretrained backbone does not. Against five open-source action-conditioned world models it leads four of the five WorldSimProbe suites and posts the best overall EWMBench score, while fine-grained local calibration still trails the field; its rollouts are faithful enough to support GRPO-based policy optimization. We release the model weights, inference code, and training code.

Figures & tables

Appendix figures & tables10 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Pelican-Sim 1.0: A General World Model Simulator for Embodied Intelligence

    Sep 10, 2026Shilong Zou, Shilin Zhang, Yingji Zhang +8World ModelsAction-Conditioned World Models

  2. WorldLine: Action-Driven Visual Simulation for Robotic Manipulation

    Sep 29, 2026Shenghe Zheng, Wenbo Li, Jiyao Zhang +4Action-Conditioned Video GenerationWorld Models for Robotics

  3. Interactive World Simulator for Robot Policy Training and Evaluation

    Mar 9, 2026Yixuan Wang, Rhythm Syed, Fangyu Wu +7Interactive World ModelsWorld Model Learning