StrideWorld: Benchmarking Scene Rigidity and Action Consistency in Interactive Video World Model

Hugging Face Daily Papers Papers

Summary

StrideWorld is a benchmark for interactive video world models that fixes reference geometry and action strides before motion, revealing that most models deform scenes and fail on composed routes, while only depth-scaled command models preserve expected positions.

Interactive video world models turn user commands into the exploration of a generated scene, which requires the scene to keep its 3D structure as the camera moves and each action to keep a stable effect. Existing benchmarks mostly evaluate these properties against references fitted to the video under test, either a scene reconstructed from its own frames or a camera trajectory aligned to the command path with a free scale. Such references can absorb gradual scene deformation and changes in pace and route length, so these errors can pass undetected. We therefore introduce StrideWorld, a benchmark whose references are fixed before the motion they judge. It evaluates scene rigidity against the ground-truth geometry of the start view, and action consistency against the stride of each action, the camera motion of one command. The stride is measured on the model's first steps and then used to test the steps that follow, so each model is judged only against its own steps. Its measurements are validated on rendered videos with known camera motion, and its scores agree with human judgments. Across videos generated by open models from 140 start views, all models but one deform their scenes as the camera moves, with offsets up to 7.3 times those of a rigid rendering along the same camera path. Most models keep a steady pace when a single action is held. Along composed routes, several models often stall, and most completed routes end away from the positions predicted by each model's own strides. Only the two models that scale commands by the estimated scene depth bring most routes close to these positions. Fine-tuning two models on a few rendered clips removes most of their mid-route stalls. Code, data, and a leaderboard are publicly available at https://github.com/chuangwang123/strideworld.
Original Article
View Cached Full Text

Cached at: 11/19/26, 03:38 PM

Paper page - StrideWorld: Benchmarking Scene Rigidity and Action Consistency in Interactive Video World Model

Source: https://huggingface.co/papers/2611.24408 Authors:

,

,

,

,

,

,

,

,

,

,

Abstract

StrideWorld is a benchmark that fixes references before motion to detect scene deformation and action inconsistency in interactive video world models, showing that most models deform scenes and only depth-scaled command models preserve composed routes.

Interactive video world modelsturn user commands into the exploration of agenerated scene, which requires the scene to keep its3D structureas the camera moves and each action to keep a stable effect. Existing benchmarks mostly evaluate these properties against references fitted to the video under test, either a scene reconstructed from its own frames or acamera trajectoryaligned to the command path with a free scale. Such references can absorb gradual scene deformation and changes in pace and route length, so these errors can pass undetected. We therefore introduceStrideWorld, a benchmark whose references are fixed before the motion they judge. It evaluatesscene rigidityagainst theground-truth geometryof thestart view, andaction consistencyagainst thestrideof each action, thecamera motionof one command. Thestrideis measured on the model’s first steps and then used to test the steps that follow, so each model is judged only against its own steps. Its measurements are validated on rendered videos with knowncamera motion, and its scores agree with human judgments. Across videos generated by open models from 140start views, all models but one deform their scenes as the camera moves, with offsets up to 7.3 times those of arigid renderingalong the samecamera path. Most models keep a steady pace when a single action is held. Alongcomposed routes, several models often stall, and most completed routes end away from the positions predicted by each model’s ownstrides. Only the two models that scale commands by theestimated scene depthbring most routes close to these positions.Fine-tuningtwo models on a few rendered clips removes most of theirmid-route stalls. Code, data, and a leaderboard are publicly available at https://github.com/chuangwang123/strideworld.

View arXiv pageView PDFProject pageGitHub2Add to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2611.24408 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2611.24408 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2611.24408 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

PlayWorld: Benchmarking World Models with Agent Players over Long-Horizon Objectives

Hugging Face Daily Papers

PlayWorld is a benchmark for evaluating interactive video world models using multi-modal agent players pursuing long-horizon objectives. It assesses geometry consistency, interaction fidelity, and state evolution, revealing that current models struggle with spatial consistency and persistent state evolution.

HappyWorld-Bench

Hugging Face Daily Papers

HappyWorld-Bench is a comprehensive benchmark that evaluates the reliability of world models under interaction and modification across video, spatial, and embodied tracks.

Dream4ACT: A Shared Visual Action Interface for Multi-Embodiment Video-Action Modeling

Hugging Face Daily Papers

Dream4ACT introduces a shared visual action interface ("action views") rendered via URDF-based forward kinematics to enable joint video-action modeling across multiple robot embodiments, combining observations and actions in a single diffusion world model with masked flow matching. It achieves 88.98% success on RoboTwin 2.0 and 65.66 on TriWorldBench using a training-free URDF-constrained recovery mechanism for executable actions.