ABot-PhysWorld: Interactive World Foundation Model for Robotic Manipulation with Physics Alignment

2026-03-24

AI Summary

This work addresses the prevalent issue of physically implausible behaviors—such as object interpenetration and anti-gravity motion—in existing video world models for robotic manipulation tasks. To mitigate this, we propose a physics-aligned video generation paradigm featuring a 14B-parameter diffusion Transformer trained on 3 million physically annotated manipulation videos. Our approach integrates a novel DPO-based post-training framework, a decoupled discriminator, and a parallel context block for precise action conditioning. We further introduce EZSbench, the first training-agnostic, zero-shot embodied evaluation benchmark. The proposed method achieves state-of-the-art performance on both PBench and EZSbench, significantly outperforming Veo 3.1 and Sora v2 Pro in physical plausibility and trajectory consistency. EZSbench is publicly released to foster standardized evaluation in embodied video generation.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Intelligent Robots: Embodied AI
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Video-based world models offer a powerful paradigm for embodied simulation and planning, yet state-of-the-art models often generate physically implausible manipulations - such as object penetration and anti-gravity motion - due to training on generic visual data and likelihood-based objectives that ignore physical laws. We present ABot-PhysWorld, a 14B Diffusion Transformer model that generates visually realistic, physically plausible, and action-controllable videos. Built on a curated dataset of three million manipulation clips with physics-aware annotation, it uses a novel DPO-based post-training framework with decoupled discriminators to suppress unphysical behaviors while preserving visual quality. A parallel context block enables precise spatial action injection for cross-embodiment control. To better evaluate generalization, we introduce EZSbench, the first training-independent embodied zero-shot benchmark combining real and synthetic unseen robot-task-scene combinations. It employs a decoupled protocol to separately assess physical realism and action alignment. ABot-PhysWorld achieves new state-of-the-art performance on PBench and EZSbench, surpassing Veo 3.1 and Sora v2 Pro in physical plausibility and trajectory consistency. We will release EZSbench to promote standardized evaluation in embodied video generation.

Problem

• physically implausible manipulation
• video-based world models
• embodied simulation
• physics alignment
• robotic manipulation

Innovation

• Physics-aware video generation
• Diffusion Transformer
• DPO-based post-training
• Embodied zero-shot benchmark
• Action-controllable world model