Paper 2603.22212
Omni-WorldBench: Towards a Comprehensive Interaction-Centric Evaluation for World Models
- Published
- Mar 2026
- Research lab
- Independent
- Citations
- 5
- GitHub
- 107 stars
01 In brief
Summary
Omni-WorldBench is a new benchmark for evaluating the interactive response capabilities of video-based world models, addressing the gap left by existing benchmarks that focus on visual fidelity or static 3D reconstruction.
It comprises Omni-WorldSuite, a set of 1,068 prompts with initial frames and optional camera trajectories, organized into three interaction levels (self, local, global) and covering general and task-oriented scenes (driving, robotics, gaming).
Omni-Metric, an agent-based evaluation framework, measures generated video quality, camera-object controllability, and interaction effect fidelity (including long-term stability, non-target stability, causal coverage, and event ordering), aggregating into an AgenticScore.
Evaluations of 18 models (T2V, I2V, camera-conditioned) show that I2V models like Wan2.2 achieve the highest AgenticScore (75.92%), while camera-conditioned models like WonderWorld excel in camera control but struggle with interaction stability.
The benchmark reveals that current models are strong in visual quality but limited in causally consistent interaction dynamics, highlighting the need for further research in 4D world modeling.
Omni-WorldBench will be publicly released to foster progress in this area.
02 From the paper
Abstract
Video--based world models have emerged along two dominant paradigms: video generation and 3D reconstruction. However, existing evaluation benchmarks either focus narrowly on visual fidelity and text--video alignment for generative models, or rely on static 3D reconstruction metrics that fundamentally neglect temporal dynamics. We argue that the future of world modeling lies in 4D generation, which jointly models spatial structure and temporal evolution. In this paradigm, the core capability is interactive response: the ability to faithfully reflect how interaction actions drive state transitions across space and time. Yet no existing benchmark systematically evaluates this critical dimension. To address this gap, we propose Omni--WorldBench, a comprehensive benchmark specifically designed to evaluate the interactive response capabilities of world models in 4D settings. Omni--WorldBench comprises two key components: Omni--WorldSuite, a systematic prompt suite spanning diverse interaction levels and scene types; and Omni--Metrics, an agent-based evaluation framework that quantifies world modeling capabilities by measuring the causal impact of interaction actions on both final outcomes and intermediate state evolution trajectories. We conduct extensive evaluations of 18 representative world models across multiple paradigms. Our analysis reveals critical limitations of current world models in interactive response, providing actionable insights for future research. Omni-WorldBench will be publicly released to foster progress in interactive 4D world modeling.