WorldGuide: Goal-Directed Video World Model for Procedural Task Execution

Hugging Face Daily Papers Papers
video-generation world-model procedural-tasks closed-loop-execution planner-executor benchmark hierarchical-memory

Summary

提出了 WorldGuide，一种面向长程程序化任务的闭环视频世界模型，通过规划器与执行器耦合及分层视觉记忆，在自建的 WorldGuide-Bench 上以 33.33% 任务成功率超过 MiniMax-H3。

Video generators and video-based world models can synthesize plausible visual trajectories, but long-horizon procedural tasks require generation to adapt to what has actually been produced. A model must determine the next action from its generated state, execute that action, and recognize when the task is complete. Open-loop generation cannot adapt to execution outcomes, while existing closed-loop systems often rely on pretrained executors or indirect verification. This leaves a gap between deciding an action and successfully realizing it. We formulate procedural video generation as closed-loop task execution in visual world space and introduce WorldGuide. Given only an initial image and a task goal, WorldGuide predicts an atomic action, generates its corresponding video clip, and uses the generated result to select the next action or terminate. The Planner and Executor are trained on the same step-level procedural demonstrations: the Planner learns to predict the next atomic action or task completion from visual progress, while the Executor is directly trained to realize the predicted actions. Hierarchical visual memory maintains state across long-horizon execution with bounded history token cost. Due to the lack of step-level action-video supervision for joint planner-executor training, we introduce WorldGuide Bench: approximately 59K step-annotated videos across 245 tasks and 27 procedural categories. WorldGuide achieves a 33.33\% Task Success on WorldGuide-Bench, compared with 29.90\% for the strong recent video model MiniMax-H3, even though MiniMax-H3 receives reference action plans, and achieves 47.69\% on VideoCraft-Bench compared with 32.73\% for MiniMax-H3 under goal-only conditioning. These results demonstrate the importance of coupling planning with learned execution for goal-directed procedural video generation.
Original Article
View Cached Full Text

Cached at: 10/09/26, 10:20 AM

Paper page - WorldGuide: Goal-Directed Video World Model for Procedural Task Execution

Source: https://huggingface.co/papers/2610.12459

Abstract

Videogeneratorsandvideo-basedworldmodelscansynthesizeplausiblevisualtrajectories,butlong-horizonproceduraltasksrequiregenerationtoadapttowhathasactuallybeenproduced.Amodelmustdeterminethenextactionfromitsgeneratedstate,executethataction,andrecognizewhenthetaskiscomplete.Open-loopgenerationcannotadapttoexecutionoutcomes,whileexistingclosed-loopsystemsoftenrelyonpretrainedexecutorsorindirectverification.Thisleavesagapbetweendecidinganactionandsuccessfullyrealizingit.Weformulateproceduralvideogenerationasclosed-looptaskexecutioninvisualworldspaceandintroduceWorldGuide.Givenonlyaninitialimageandataskgoal,WorldGuidepredictsanatomicaction,generatesitscorrespondingvideoclip,andusesthegeneratedresulttoselectthenextactionorterminate.ThePlannerandExecutoraretrainedonthesamestep-levelproceduraldemonstrations:thePlannerlearnstopredictthenextatomicactionortaskcompletionfromvisualprogress,whiletheExecutorisdirectlytrainedtorealizethepredictedactions.Hierarchicalvisualmemorymaintainsstateacrosslong-horizonexecutionwithboundedhistorytokencost.Duetothelackofstep-levelaction-videosupervisionforjointplanner-executortraining,weintroduceWorldGuideBench:approximately59Kstep-annotatedvideosacross245tasksand27proceduralcategories.WorldGuideachievesa33.33\%TaskSuccessonWorldGuide-Bench,comparedwith29.90\%forthestrongrecentvideomodelMiniMax-H3,eventhoughMiniMax-H3receivesreferenceactionplans,andachieves47.69\%onVideoCraft-Benchcomparedwith32.73\%forMiniMax-H3undergoal-onlyconditioning.Theseresultsdemonstratetheimportanceofcouplingplanningwithlearnedexecutionforgoal-directedproceduralvideogeneration.

View arXiv pageView PDFProject pageGitHub6Add to collection

Get this paper in your agent:

hf papers read 2610\.12459

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper1

#### ankanmbz/WorldGuide-Ckpt Image-to-Video• 9B• Updatedabout 4 hours ago

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.12459 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.12459 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

ExeWorldBench: Benchmarking Code-Based Executable 3D World Modeling

Hugging Face Daily Papers

ExeWorldBench 是首个通过基于代码的 3D 世界构建，全面评估世界理解与可执行实现能力的基准，包含三个递进任务：机制可执行性、组合可执行性和意图驱动的可执行世界合成。实验显示前沿模型在可执行世界建模方面仍存在显著局限。

Long-WAM: Scaling the Context of World-Action Models

Hugging Face Daily Papers

Long-WAM 提出了一种在实时控制约束下扩展因果世界-动作模型上下文的模型-系统框架，发现更长的视觉历史仅在视频基础模型以自回归方式预训练时才显著有效，在 RoboCasa GR-1 上将成功率从 63.3% 提升至 78.7%，并在 Unitree G1 实机上实现 95% 的动态杯子堆叠成功率。

GameWAM: A World Action Model for Video Games

arXiv cs.AI

GameWAM introduces the first world-action model for native closed-loop gameplay and GUI control in video games, jointly generating visual observations and executable actions with competitive task success and revealing a source-sensitivity failure mode.