World Model Evaluation

Follow this topic Ask about this topic

Momentum

29 papers in the last four weeks, up 123% on the four weeks before. 0.3% of all new papers.

Jul 13Week of Sep 28

Latest papers 146

Oct 7, 2026cs.AI

RoboJEPA: Scaling Robotic Latent World Models

Latent world models have shown a remarkable ability to predict future states and to plan in the real world. In practice, however, we lack a principled way to estimate how their capabilities scale with model size, data, and compute, an open problem that slows progress in the field. In this work we present RoboJEPA, a world model based on the Joint Embedding Predictive Architecture (JEPA) and trained on a large-scale dataset spanning 12 robotic embodiments. We show that RoboJEPA's imagination error, the error of its latent rollouts, follows a second-order power law in compute, allowing us to predict model quality well beyond the scale at which the law is fit. We further show that downstream robotic planning performance improves predictably with compute, and that imagination error is strongly correlated with it, making it a reliable proxy for real-robot evaluation. Finally, we demonstrate that latent world models can be deployed zero-shot as robotic agents, planning toward a single goal image to solve tasks requiring long-horizon planning on real hardware. We release all model checkpoints together with our training and robot deployment code. To our knowledge, this is the first work to establish scaling laws for multi-embodiment robotic world models trained on real robot data, and RoboJEPA, at 8B parameters, is the largest JEPA predictor model trained to date.

Artem Zholus, Nicolas Beltran-Velez, Jianhao Yuan +9

FAIR at Meta · Chandar Research Lab · Mila - Quebec AI Institute +2

World Models Latent World Models World Model-Based Planning World Model Evaluation Joint-Embedding Predictive Architecture Cross-Embodiment Robot Learning

Oct 7, 2026cs.LG

Kuration SDK: Addressing the Virtual2Real Gap via Data Curation

Benchmarks for measuring the quality of action-conditioned world models are still evolving and shifting away from visual similarity-based metrics to action-semantic and physically-grounded metrics. However, for domain and task-agnostic action-conditioned world model training, existing benchmarks provide a limited signal. By training and evaluating diffusion world models on CounterStrike gameplay data, we confirm that qualitative playability does not correspond with metrics such as FVD, LPIPS, and JEDi. We term this the Virtual2Real gap. We posit that, in lieu of reliable benchmarks, curating raw gameplay data and measuring a variety of diagnostic properties provides a more robust signal to bridge the gap, before the training even begins. We present several curation strategies and a general-purpose kit for physical AI data curation called Kuration SDK, which is being open-sourced with this paper. The SDK was instrumental in uncovering the root cause of the virtual2real gap in a specific case: why two world models trained on identical gameplay map, action and state distribution, behaved very differently when played in spite of having very similar LPIPS and FVD scores. Thus, Kuration SDK has the potential to uncover the root causes of Virtual2Real gap in specific datasets and accelerate development of sample-efficient training datasets.

Nirmit Desai, Eric Song, Mayank Sengupta +4

AIntropy AI

Training Data Curation Action-Conditioned World Models World Model Evaluation Video Diffusion Models

Oct 6, 2026cs.RO

World Models Dream of Success: Diagnosing and Repairing Failure Insensitivity in Robot World Models

Robot world models support policy evaluation, planning, and synthetic data generation, but these applications require predictions that distinguish successful actions from failures. Across four released checkpoints from two architecture families, we observe weak sensitivity to action changes and success-like predictions on verified failures. Although recent work incorporates failures into model training, which data can repair released checkpoints without changing their architecture or training objective still remains underexplored. To this end, we introduce CureWM, which constructs alternative actions from successful demonstrations across a severity grid, verifies their outcomes through execution in simulation or on hardware, and fine-tunes released models on the resulting failures and surviving successes alongside nominal demonstrations. This construction provides controlled action contrasts from shared starting contexts. On 484 held-out LIBERO failure counterfactuals, optimism falls from 80% after fine-tuning on the official data to 30--43% across four independently fine-tuned CureWM models (38% mean). In two separate evaluations on a physical robot arm, failure predictions scored as success-like by a latent-distance diagnostic decrease from 90% after fine-tuning on successful demonstrations alone to 33% with CureWM. With failure counts per task, successful replay data, and training budget matched, counterfactual failures yield a success--failure value gap of 0.124, compared with 0.014 for freshly collected on-policy failures. These findings support execution-verified counterfactual replay for post-hoc repair and show why reduced optimism must be evaluated alongside success--failure discrimination. Code is available at https://github.com/jiuyixu25/CureWM.

Jiuyi Xu, Xiao Hu, Meida Chen +3

Colorado School of Mines · Northeastern University · Institute for Creative Technologies, University of Southern California +1

World Models for Robotics Action-Conditioned World Models World Model Evaluation Counterfactual Learning

Oct 6, 2026cs.CV

World Models' Last Exam in Physics

Video world models can produce visually convincing yet physically inconsistent sequences, raising concerns about their reliability for prediction and planning in embodied AI systems. Existing evaluations often rely on model-based judgments or reference videos, while direct physical tests largely focus on mechanics. We introduce World Models' Last Exam in Physics, a measurement-based benchmark for evaluating physical consistency in video world models. The benchmark comprises 40 controlled tasks spanning mechanics, optics, fluids, thermal and phase-change phenomena, electromagnetism, and surface tension. Each task pairs an initial image and a generation prompt with predefined physical criteria, enabling interpretable tests of observable physical relationships without requiring reference videos. Its evaluator combines task-observability screening with task-specific quantitative physical measurements. Experiments on eight video generation models across 1,280 videos reveal persistent physical inconsistencies and substantial variation across tasks, with the best model achieving an overall score of 57.76 out of 100. Evaluation on synthetic videos with known physical relationships provides evidence for the validity of the measurement module under controlled conditions. The evaluator also achieves higher agreement with human judgments than a direct vision-language model baseline in both within-task rankings and pairwise comparisons. By combining coverage across physical domains with scores grounded in measurable evidence and explicit measurement limitations, the benchmark provides an interpretable basis for diagnosing physical inconsistencies and tracking progress toward physically consistent video world models.

Mingju Gao, Qingle Liu, Yuzhao Peng +8

Navers Lab, Einsia.AI · Peking University · Tsinghua University

Physical Consistency in Video Generation World Model Evaluation Physical Reasoning in Video Video World Models

Oct 5, 2026cs.CV

Tracking Is Not Permanence: What Video World Models Keep of a Hidden Object

Video world models track objects they can see; we ask what they keep of objects they cannot. We hide an object from a frozen V-JEPA 2 predictor and compare its prediction for the hidden region with the encoder's representation of two worlds that differ only inside that region. The predictor's decision keeps a stationary object in part and one carried inside a container not at all, and loses a moving one within 0.3 s (0.5 s under V-JEPA's own tube mask; ViT-H keeps it to 1.1 s at pretraining's 90% masking ratio); in projection a trace remains, below the midpoint, at 14-60% of what a baseline copying the last view retains. The information is there: the encoder reads the object's presence at 1.00 and keeps a closed container's contents decodable for 3.5 s, while the predictor's output, read with the encoder's own probe, contains the ball in 2% of scenes once the box has been closed for half a second. On rendered scenes, permanence is missing on the predictor's side, and training installs it cheaply as a prior: three thousand predictor-only steps on synthetic containers take this belief from 0.05 to 1.00 against two matched controls. They also raise IntPhys-2019 from 84.2% to 93.3%, but so does a curriculum without containers, and which training habit the benchmark credits changes with its scoring rule. Continued training with tube masks produces 1.1-1.6 s of moving-object carry-over on manipulation and internet-style video, so the deficit is not intrinsic to latent prediction. VideoMAE keeps almost nothing, and Cosmos's next-token prediction keeps a stationary hidden object but not one carried inside a moving container.

Peng Xie, Amr Alanwar

TUM School of Computation, Information and Technology Technical University of Munich, Germany

Temporal Consistency Object Permanence World Model Evaluation Video World Models

Oct 4, 2026cs.LG

When Low Prediction Error Misleads Planning: Diagnosing Representation, Dynamics, and Decision Failures in Latent World Models

The component that dominates a latent world model's prediction error need not be the one whose repair most improves action selection. We show this by comparing action sequences from identical physical starts and separating endpoint error into a candidate-pool center and action-relative responses. Across four model families and four tasks, a confirmation pool of 256 new starts per task and 300 shared candidates per start shows that center error dominates MSE in 14/16 model-task cells. Yet in six of these cells, an oracle that corrects only the action-relative responses yields better physical rank correlation and top-30 elite quality than one that corrects only the center, while leaving more latent MSE (family-wise corrected intervals). The preference differs across the evaluated settings: a separate LeWorldModel (LeWM) study that executes oracle-selected actions favors center repair on PushT and on Reacher with a render-matched goal. Matched-candidate tests localize ordering loss: for LeWM, encoding realized endpoints raises physical Spearman from 0.464 to 0.975 on that Reacher setting and from 0.193 to 0.631 on PushT (64 starts per task), while Cube's encoded-goal cost remains uninformative. A 72-run objective study improves selected response diagnostics, while incremental closed-loop planning gains remain unconfirmed. These results separate error magnitude from the decision effects of oracle correction and motivate evaluating representation, prediction, and planning as separate stages.

Rui Min, Xianyao Li, Fang Xu +1

Department of Mechanical and Aerospace Engineering, University of Florida, Gainesville, FL 32611, USA · Department of Civil and Coastal Engineering, University of Florida, Gainesville, FL 32611, USA

World Model-Based Planning Action-Conditioned World Models Latent World Models World Model Evaluation

Oct 3, 2026cs.CV

Video2World: Benchmarking Coding Agents for Interactive World Modeling from Embodied Videos

Building interactive simulators from real-world observations is a promising way to scale embodied data, but current pipelines still rely heavily on manual environment construction and calibration. We study whether frontier foundation models and coding agents can automate this process end to end. We formulate \emph{autonomous video-to-simulation} as a software engineering task in which an agent observes an embodied video, constructs the corresponding simulated environment and robot behavior, and iteratively refines the result through execution feedback. To evaluate this capability, we introduce \textbf{Video2World}, a benchmark comprising 222 reconstruction instances derived from 189 robot and human demonstration videos. Video2World measures reconstructed worlds along geometric fidelity, dynamic fidelity, and functional correctness, capturing spatial perception, physical reasoning, and executable interaction. Evaluating 9 frontier coding-agent systems reveals a sharp improvement in Task success beginning with Claude Opus 5, rising from below 5% to over 15%, while substantial gaps to human-assisted reconstruction remain. We further find that worlds that look better could work worse: better visual fidelity does not always lead to higher task success. This echoes the broader gap between perceptual realism and factual correctness observed in generative models.

Jinzhou Tang, Zijun Zhang, Jing Yang +10

Aether AI · University of California, San Diego

Interactive World Models Coding Agents World Model Evaluation AI Agent Benchmarks

Oct 1, 2026cs.CV

PROWBench: Do Video Models Render What the Program Specifies?

Programmable world models separate executable dynamics from visual generation, offering a promising foundation for next-generation game engines. However, their visual adherence to explicit rules and interactions remains insufficiently evaluated. Existing benchmarks assess visual quality, controllability, and instruction or physical adherence, but rarely test fidelity to fine-grained, program-specified world events. We introduce PROWBench, comprising 170 programmatically constructed episodes and 600 proxy videos covering diverse scenes and interactions. PROWBench logs entity states and timestamped events, including those outside the camera's field of view, as replayable world records, from which it renders synchronized views and proxy representations. This enables generated videos to be checked against the observable consequences of program execution. An extensible framework constructs scenes, controls behaviors, and can render each camera view in different representations, such as coarse 3D, and bounding boxes. The benchmark covers first- and third-person perspectives, with synchronized multi-view observations available for a subset of episodes. Grounded in these records, PROWBench evaluates entity control, long-horizon memory, and, with two VLM-based metrics, Logic-Render Alignment and Interaction Success Rate, adherence to the prescribed timeline and the visual realization of timestamped engine-recorded events.

Zheng-Hui Huang, Guixu Lin, Yu-Ju Tsai +6

Alaya Lab

Long-Horizon Video Generation Controllable Video Generation World Model Evaluation Video World Models Video Generation Evaluation

Oct 1, 2026cs.LG

In CEM, a World Model Is Also a Proposal Mechanism

The cross-entropy method (CEM) uses world-model scores to select action sequences and fit the distribution sampled in its next iteration. A scoring error can therefore change both the present decision and the candidates considered later. We evaluate these two roles separately. Four types of predictive model generate CEM traces, and every model rescores every saved candidate pool. Executing the same candidates in the environment provides a reference elite set and proposal update. Across twelve independently trained task-seed units on Walker and Cheetah, the pre-specified proposal distance falls from the first to the final CEM iteration in every unit. Proposal widths contract and fitted means separate relative to the remaining search width. Pairwise ranking agreement stays near chance on Walker and declines on Cheetah; elite-set agreement does not improve. This comparison shows greater variation between scorers than between pool sources on Cheetah; Walker has variation in both and in their pairings. We use the original six units to select Random nonlinear for a one-update intervention, without inspecting intervention outcomes. Replacing its first model-ranked update with an environment-ranked update lowers final realised selected-sequence cost in those six units and in six further units held out from the selection.

Oliver Obst, Frieder Stolzenburg

UNSW Sydney · Harz University of Applied Sciences

Black-Box Optimization Model-Based Planning World Model Evaluation Model-Based RL

Sep 30, 2026cs.AI

Kepler: Auditable World Models for ARC-AGI-3

ARC-AGI-3 evaluates agents in interactive environments whose rules and objectives must be inferred from observation. We present Kepler, an open-source harness that represents hypotheses as executable world models and validates them through retrospective transition checks and conditional prediction checks. Under one frozen Claude Opus 5 configuration, Kepler obtained a server-verified 100.00 RHAE on all 25 public games, with no per-game model selection or score-conditioned reruns. On 181 of 183 completed levels, the final Opus attempt used no more actions than the corresponding median-human baseline. The retained board runs used 8,256 environment actions, of which 7,292 occurred in scored levels. Retained local provider-session records yield 858.0 million tokens, 97.37% cache reads, and a $777.72 cost at September 1, 2026 API list-equivalent rates. We also report three evaluation failures: source-code leakage that produced an invalid perfect run, agents reconstructing a removed harness in a control condition, and autonomous repair masking a broken planner. A single-game observation case study showed that animation frames contained task-relevant information absent from settled text grids. Across the final Claude Opus 5 and GPT-5.6 Sol boards, 48 of 50 game-model cells reached 100. These results indicate that public-set score alone has limited discriminative value and motivate first-attempt, cost-conditioned, and verification-aware reporting.

Wensen Wu

Independent Researcher

Long-Horizon Agent Evaluation ARC-AGI AI Agent Benchmarks World Model Evaluation

Sep 30, 2026cs.AI

Why Do Conventional World Models Fail to Learn Cellular Automata?

Although conventional world models - auto-regressive or diffusion models based on transformers or convolutional networks - may learn surface statistics of world dynamics, can they learn the exact world dynamics from its observed history? Leveraging cellular automata as a simple testbed, we find the answer to be no in many cases. Conventional architectures predict most pixels correctly yet rarely complete a rollout: a CNN predicts 96.3% of cells but completes 18.9% of rollouts; a joint diffusion model completes none. We trace the gap to three failure modes of these world models - namely, they fail to exactly capture spatial locality, temporal locality or temporal stability. Simple changes repair each: (1) for spatial locality, two-dimensional rotary positions lift a transformer from 39.1% to 100% on the Game of Life; (2) for temporal locality, handing each token its cell's previous-frame neighbourhood lifts the same transformer from 25.8% to 99.9% on unseen rules; (3) for temporal stability, causal freezing lifts the same diffusion weights from 42.2% to 99.9%. None of the three changes touches the architectural backbone; each only modifies the information flow within it. We also compare joint and ordered sampling on billiards and, in an exploratory study, on a simulated Burgers equation.

Shaoyang Guo, Ziming Liu

Meta Circle(元环智能) · Peking University · Tsinghua University

World Model Learning Spatiotemporal Reasoning World Model Evaluation Diffusion Models

Sep 30, 2026cs.RO

The Planning Limits of Latent World Models

World models offer a promising way to help robots understand how the physical world evolves and plan complex behaviours through imagination. Yet existing studies mainly demonstrate what these models can accomplish, leaving unclear when their predictions remain useful for planning and where they fail. We study this question using action-conditioned predictors built on five frozen self-supervised visual backbones: V-JEPA 2, V-JEPA 2.1, VideoMAEv2, VideoPrism, and DINOv2. We use frozen backbones to test representations intended to transfer across environments. We evaluate these models on diverse Meta-World manipulation tasks and real-robot interactions from BridgeData V2. We find that a world model guides action selection reliably only when the goal lies within, or slightly beyond, the trajectory it imagines during planning. With five-step rollouts, the length the predictor was trained on, the world model ranks actions reliably only for targets five to ten control steps ahead, whereas task goals lie 16 to 53 steps away. Neither an 81-fold larger predictor nor longer-rollout training extends this range; the encoder affects both range and closed-loop success, with V-JEPA 2.1 performing most consistently. More fundamentally, the limit persists under perfect prediction: using the real simulator, success falls from 92% to 41% as the target moves from five to twenty steps ahead of a five-step rollout. Planning therefore requires either longer imagined trajectories or closer subgoals. For distant goals, pure imagination succeeds in 23% of episodes, planning with feedback (MPC) raises success to 30%, imagining as far as the goal to 47%, and nearby expert subgoals to 76%. Used within its plannable range, a world model can also improve a vision-language-action (VLA) policy: choosing among eight actions the VLA proposes raises its success from 65% to 77% across 16 different tasks.

Ali Alrasheed, Basim Azam, Naveed Akhtar

The University of Melbourne, Australia

World Model-Based Planning Action-Conditioned World Models Robotic Manipulation Model-Based Planning Latent World Models World Model Evaluation

Sep 29, 2026cs.CV

HelixWorld: A Real-time Interactive Audio-Visual World Model

World simulation is inherently multisensory, demanding synchronized visual and acoustic dynamics in real time. Yet prevailing interactive world models remain strictly silent, focusing exclusively on visual rendering and control while overlooking the acoustic dimension. We present HelixWorld, a real-time interactive audio-visual world model where visual scenes and camera-grounded spatial stereo sound co-evolve natively under user interaction. We curate a high-fidelity spatial audio-visual dataset with true stereo acoustics and metric camera poses, upon which we pre-train a bidirectional teacher conditioned on 6-DoF camera trajectories and user actions. To enable low-latency causal interaction, we distill the teacher into a few-step streaming student via an online trajectory distillation loss, sustaining drift-free joint audio-visual rollouts at 24 FPS on a single GPU. Furthermore, we formalize spatial-acoustic consistency and introduce HelixBench to evaluate whether synthesized sound fields faithfully track dynamic viewpoint motion. Extensive experiments demonstrate that HelixWorld matches state-of-the-art silent world models in visual fidelity and responsiveness, while significantly surpassing existing baselines in camera-aligned spatial-acoustic immersion.

Lei Ke, Jiahao Pan, Zeyue Tian +13

The Hong Kong University of Science and Technology · Noiz AI

World Model Learning Real-Time Interactive Video Generation Audio-Visual Synchronization Spatial Audio World Model Evaluation Interactive World Models

Sep 29, 2026cs.CV

Foresight at the Event Boundary: Evaluating Physical Prediction in Video World Models

Video world models are largely regarded as predictive models of the physical world and are therefore expected to anticipate the consequences of observed events. However, evaluation has mainly focused on reference similarity, physical-law consistency, or judgment plausibility, estimating anticipation only indirectly. We address this directly: when a release or impact has just occurred but its consequence is withheld, can a world model anticipate what should happen next? We introduce an event-anchored evaluation based on 62 controlled real-world free-fall recordings and 124 clips spanning three object types, with fine-grained release and impact annotations and ground-truth trajectories. The protocol separates consequence production, temporal placement, and physical realization. Across six contemporary video generation and world models, Runway and Veo produce release and subsequent impact events at rates above 93% but often initiate them substantially late, whereas Cosmos-Predict-2.5 and MAGI-1 frequently preserve the pre-event state and produce little or no measurable consequence. Among measurable falls, plausible timing does not necessarily imply physically consistent motion. We further conduct a 15-participant, 20-condition human study in which participants describe the expected consequence from a single event-anchored frame and draw its trajectory. Human predictions favor the recorded future in aggregate while revealing genuine ambiguity among plausible continuations. Overall, physical foresight emerges as a sequence of distinct challenges: initiating a consequence, anchoring it in time, and realizing its motion.

Estela Monserrat Arriaga Santana, Julian Rosas Scull, Ehécatl Sacamch'en Núñez Rico +1

National Autonomous University of Mexico · National Autonomous University of Mexico, Mexico · University of Texas at El Paso +1

Physical Consistency in Video Generation World Model Evaluation Physical Reasoning in Video Video World Models Temporal Video Understanding

Sep 28, 2026cs.CV

OPIS: An Input-Grounded Benchmark for Multi-Object Memory in Video World Models

Video world models must preserve the visual state of the world over time, but existing evaluation protocols often rely on generated histories, video reference, or selected revisit viewpoints that can confound the assessment of a model's true memory capability. To address this, we introduce OPIS, an input-grounded benchmark that strictly anchors the assessment to a fixed set of object instances from the initial observation for evaluating multi-object memory in video world models. The OPIS dataset comprises 500 cases across real-world, embodied-robotic, and game-world domains, providing dense object-level annotations for 12,672 rigid, articulated, and deformable instances. Our object-centric evaluator combines association and explicit visibility reasoning to hierarchically measure Object (O) Presence (P), Identity (I), and Structure (S), utilizing static or dynamic evaluation tracks based on object kinematics. Across eight image-to-video or camera-conditioned world models, our proposed OPIS scores range from 48.65 to 56.01. As the reference inventory grows from less than 20 to more than 40 objects, the Presence, Identity, and Structure scores show an overall decline, with the average Identity score falling from 40.22 to 23.11. The results demonstrate that preserving the particular object instances in the input is considerably harder than generating plausible visual elements.

Hao Wang, Tao Yu, Liuzhou Zhang +13

USTC · Infrec · CASIA +10

World Model Evaluation Video World Models

Sep 27, 2026cs.LG

Behavioral Monitoring of JEPA World Models with Jacobian Centroids

Detecting failures in World Model (WM)-based planning requires monitoring whether the model is behaviorally aligned with the current task, which in turn requires studying its internal representations. Here, we show that centroids---sub-component Jacobian row-sums---effectively identify the behavioral properties of WMs, complementing traditional activation-based knowledge signals. The centroids of a model are easily computed through Jacobian vector products and characterize how the model organizes the geometry of its input space, yielding an efficient perspective on internal representations, including the generation of task-relevant saliency maps. Evaluated on continuous control tasks using JEPA WMs, this behavioral view reveals a structural dissociation, where the encoder correctly represents the goal while the predictor remains behaviorally unresponsive. This failure mode directly predicts planning failure before any action is taken, allowing for goal resampling to recapture out-of-distribution success. Moreover, centroid-based methods outperform baseline methods as distribution-shift detectors. Together, these tools yield a behavioral monitoring stack that is operational and consequential under distribution shifts.

Thomas Walker, Randall Balestriero, Richard Baraniuk

Rice University · Brown University

Distribution Shift World Model-Based Planning Joint-Embedding Predictive Architecture Distribution Shift Detection World Model Evaluation

Sep 27, 2026cs.LG

MultiEcho: An Experimental Science of Learned Worlds

World models can be studied as experimental systems with response laws of their own. We introduce MultiEcho, a framework for estimating these laws through controlled counterfactual interventions, delimiting their applicability, and separately testing their physical correspondence. Across nine simulated physical systems and seven frozen model configurations, three-reference estimators predict complete intervention responses and recover intervention parameters. Estimator selection uses discovery data only; frozen fits are evaluated on validation and confirmation contexts. The experiments distinguish response predictability, intervention readability and physical accuracy. Responses can be locally describable yet poorly match physical effects in the same target coordinates. Event-window, visibility and camera interventions reveal conditional applicability, and paired generator configurations show reduced readability under a scene prompt with stronger guidance. Magnitude sweeps expose small image errors alongside large relative effect errors. An exact-reset material experiment separates registered visible-response success from fixed-readout failure on material-dependent futures at matched positions and velocities. Exact finite-scale identities resolve odd and even response errors; first-order remainder bounds specify when refined calibration converges. MultiEcho provides an experimental basis for studying learned-world laws independently of, and in relation to, physical laws.

Meng Zhu, Airui Zhang

SHIHAO XINRUI Group

Counterfactual Evaluation World Models World Model Evaluation

Sep 27, 2026cs.CL

Does Learning to Predict the World Help Agents Act? Auditing World-Model Post-Training

Predicting how an environment will change before acting is a natural route to better decision making for agents. Recent post-training methods therefore require agents to predict the next observation and turn that prediction into a reward or a direct supervision signal, which is called world model. Existing next-observation training methods help the agent to learn the environmental content. However, they additionally involve an optimization process, which may introduce several effects other than learning to predict the world. Consequently, where the performance gain comes from during the training process remains an open question. We answer this research question through replacing true next-observation targets with in-distribution mismatched observations during the training process. Across two interactive text environments, mismatched targets lower prediction accuracy by 15.3-61.6% relative to ground-truth targets, yet retain substantial task gains over the base model. Compared with the base model, trained models consider more candidate actions and exhibit less looping. We also introduce a setting that replaces prediction-based rewards with independent random signals. This training expands task coverage (pass@64) even when the reward carries no environment information. We also generalize this finding to VisualWebArena, where random-reward training raises pass@64 by 14.3% relative to the base model, without observation-matching rewards or an external multimodal teacher for reward construction.

Xinyu Che, Hang Yan, Yanchen Liu +5

Xi’an Jiaotong University · University of Southern California · University of the Chinese Academy of Sciences +1

World Model Learning RL Post-Training World Model Evaluation RL for LLM Agents

Sep 27, 2026cs.CV

VehDyn: A Driving World Model Benchmark for Vehicle Dynamics

Video world models are emerging as data engines, action planners, and generative simulators for autonomous driving, but existing benchmarks primarily assess visual fidelity and coarse physical plausibility, providing limited evidence on whether generated driving futures obey realistic vehicle kinematics and dynamics. This limitation is further compounded by the lack of datasets in which vehicle, road, maneuver, and speed conditions are independently controlled, and ground-truth vehicle states are recorded in synchrony with videos. We introduce VehDyn, a driving world model benchmark for vehicle dynamics. VehDyn is built on a CARLA-CarSim co-simulation platform where photorealistic rendering is coupled with a validated multi-body dynamics model, and it contains 10,080 configurations from a full factorial design over five vehicle types, four tire-road friction coefficients, three maneuvers, four target speeds, 14 scenes, and three illuminations, each paired with synchronized position, velocity, and attitude sequences. Built on this dataset, VehDyn introduces a hierarchical evaluation framework that measures trajectory alignment, kinematic consistency, and dynamic consistency, and benchmarks 12 state-of-the-art video world models. We further assess the video quality using two established protocols and correlate it with the VehDyn score. Trajectory-level metrics are nearly saturated, with ten of twelve models within 20% of ground truth, while no model reaches 92% of ground truth on dynamic consistency, and visual-quality metrics are only weakly correlated with vehicle-dynamics fidelity. DrivingWorld achieves the highest VehDyn score, followed by Cosmos 3 Nano and LTX-Video 2.5, and the VehDyn score agrees closely with human judgment. VehDyn provides a systematic foundation for developing driving world models that are physically consistent and visually realistic.

Tianyi Wang, Wangsheng Du, Jiazhou Chen +12

Autonomous Driving Benchmarks Physical Consistency in Video Generation Vehicle Dynamics World Model Evaluation Video World Models

Sep 26, 2026cs.LG

Not All Errors Matter: Decision-Relevant Prediction Error Predicts Planning Quality

World models are typically trained and evaluated by prediction error, assuming that more accurate predictions lead to better decisions. We show that this assumption can fail because models with similar total error can differ substantially in planning performance when their errors occur on different state dimensions. We introduce Decision-Relevant Prediction Error (DRPE), which measures prediction error on the state dimensions that affect decisions. We also develop an iso-error evaluation protocol that varies error allocation while keeping total error fixed. In a factored gridworld with known state relevance and a standardized planner, we evaluate 55 controlled and learned models across different error levels and allocations. Total prediction error is weakly related to planning success (Spearman

ρ=-0.25

), whereas DRPE is strongly predictive (

ρ=-0.84

;

-0.98

within the controlled family). Models with only a 1% difference in total error can differ by 60 percentage points in planning success (97% vs 37%). The relevant error also depends on the task, with model rankings reversing across tasks at the same total error. Deeper imagination further amplifies decision-relevant errors, while learned models exhibit systematic bias on rare but decision-critical events. We formalize sufficient conditions under which DRPE correctly ranks models and total prediction error cannot.

Linhao Wang, Yiyan Fan, Dongjin Huang

Shanghai Film Academy, Shanghai University

World Model-Based Planning Model-Based Planning World Model Evaluation Model-Based RL

Sep 22, 2026cs.RO

TriWorldBench: A Tri-View Consistency Perspective on Embodied World Models

Embodied world models predict the outcomes of robot actions to support learning and planning. For robots equipped with head and wrist cameras, this requires complementary views: the head view captures the overall task, while wrist views reveal local gripper-object interactions. However, evaluating these views independently cannot determine whether they describe the same action and object state. We introduce TRIWORLDBENCH, a benchmark for evaluating embodied world models through synchronized head, left-wrist, and right-wrist videos. It contains 500 episodes across 50 bimanual manipulation tasks and uses 19 metrics to assess tri-view consistency, task alignment, physical and 3D coherence, motion quality, temporal consistency, and visual quality. By combining cross-view checks with measurements tailored to each camera, the benchmark evaluates whether plausible individual videos also form a consistent prediction of the intended task. We summarize overall performance with TWB-Score and retain per-view results to identify where predictions fail. This extends world-model evaluation beyond single-view visual quality. Code, data, and metric definitions are available at https://github.com/TriWorldBench/TriWorldBench.

Xuanyi Liu, Haofeng Wang, Ruiqi Li +9

Peking University · Tsinghua University · Beihang University +3

World Model Learning Robot Manipulation Benchmarks World Model Evaluation Embodied World Models

Sep 22, 2026cs.AI

Dual-Frontier: When Can an Agent Trust Its World Model?

Learned world models are becoming essential to general-purpose agents: by predicting action consequences, they support planning and decision-making while reducing reliance on costly trial and error. This reliance creates a fundamental ambiguity: when a world-model-guided decision fails, the trajectory alone may not reveal whether the agent's decision rule or the world model caused the loss. We formalize this failure-attribution problem as a counterfactual decomposition of return loss and prove that its components are not identifiable from passive interaction, even for finite-horizon planners. This obstruction motivates Dual-Frontier, a learning principle that admits a world-model-guided decision only when its predicted advantage exceeds a certified bound on decision-relevant world-model error; otherwise, evidence is allocated to world-model verification. Action-conditioned value bounds and a closed-loop extension guarantee non-decreasing return for admitted decisions. Calibrated gates and simultaneous confidence sequences support adaptive evidence reuse, with sufficient and necessary verification bounds. Controlled learned-model experiments validate the predicted failure modes and certification behavior, while cross-backbone tool-use benchmarks instantiate the same verify-then-promote rule in realistic agent world-model pipelines, consistently improving decision quality and reliability.

Huatai Zhu, Qiang Chen, Ziqian Kou +5

Central South University · The Hong Kong University of Science and Technology · Xiangjiang Laboratory +2

World Model-Based Planning Model-Based Planning World Model Evaluation Counterfactual Learning Model-Based RL

Sep 21, 2026cs.RO

Beyond Visual Quality: A Study of Test-Time Planning with World Action Models

World action models generate actions together with visual predictions of their consequences. These paired outputs create the potential for planning by sampling multiple actions from one state, comparing their imagined outcomes, and choosing the action with the most promising predicted outcome. However, how to use imagined futures to guide action selection remains unclear. We examine this planning potential empirically. First, we estimate an oracle upper bound on selection by choosing the sampled candidate whose realised outcome is best. In a controlled same-state analysis, this choice raises success from 68.9% under uniform random selection to 79.2%. We then test selectors based on visual quality, physical consistency, and task progression as controlled interventions. Some tested selectors yield higher observed success, but the gains are uneven and the matched selectors leave much of the measured opportunity unrecovered. To investigate this gap, we examine whether sampled actions lead to different outcomes, whether these differences are visible in the predictions, and whether a score recognises them. Counterfactual branching from the same states shows that selection opportunity is concentrated in relatively few decisions in the initial candidate sets. Action spread and outcome coverage need not increase together. In a further evaluation across trajectory phases with complete action execution, the tested scores again recover little of the available improvement despite a small gain from learned value. These findings distinguish producing consequential action choices from recognising them in generated futures, motivating the evaluation of WAM predictions through their usefulness for decisions rather than visual quality alone.

Jianhao Yuan, Yu Yuan, Benjamin Ramtoula +5

University of Oxford · Purdue University · UWE Bristol

World Model-Based Planning Action-Conditioned World Models Model-Based Planning World Model Evaluation

Sep 21, 2026cs.CV

HappyWorld-Bench

Evaluating world models requires assessing both the quality of the worlds they generate and their consistency and responsiveness under exploration, interaction, and modification. We introduce HappyWorld-Bench, a comprehensive benchmark that evaluates whether generated worlds remain reliable as agents interact with them. Our design is built on a hierarchical capability framework of six world capabilities (W1-W6), from generative construction to unified world modeling, instantiated across three independent evaluation tracks: video world models, spatial world models, and embodied world models. HappyWorld-Bench comprises 1,138 video prompts, 300 spatial scenes, and 254 embodied test cases. Across all three tracks, we build and operate HappyWorld-Arena to organize human A/B comparisons and derive model-level Elo ratings, which complement newly designed automated metrics that capture behavioral correctness. We evaluate 14 video world models, 9 spatial systems, and 8 embodied candidates under this unified framework. Results reveal remaining reliability gaps across all three tracks: video models exhibit reduced consistency during extended rollouts and revisits, spatial models achieve at best 70.14% placement accuracy and 73.33% edit execution, and embodied models struggle to preserve state across multi-step actions and respond precisely to altered action conditions and physical rules. These findings highlight the need to evaluate world models not only by visual quality, but also by state consistency and the correctness of their responses to actions and interventions.

Zhiqi Bai, Junai Cai, Yixin Chen +33

Alibaba Token Hub, Alibaba Group · State Key Laboratory of General Artificial Intelligence, BIGAI · Tsinghua University +2

World Models World Model Evaluation Interactive World Models Embodied World Models Video World Models

Sep 21, 2026cs.RO

What Matters in Designing World Action Models: An Empirical Study

World Action Models (WAMs) have emerged as a promising paradigm for generalizable robot control. Despite the growing number of WAM systems, existing works often introduce unified systems that bundle together multiple design choices, such as architecture and training strategy, making it difficult to isolate individual contributions and systematically compare alternative designs. In this work, we present a controlled study that disentangles these design choices and analyzes not only their empirical effects, but also how and why they shape WAMs. More specifically, we focus on three fundamental questions in building WAMs: (1) what causal structure should govern the interaction between world modeling and action generation? (2) in which latent space should world modeling be performed? and (3) how do different world-action modeling objectives affect model behavior and performance? Through structurally controlled experiments on three representative benchmarks, RoboCasa-GR1, LIBERO, and LIBERO-Plus, we systematically compare six causal structures, eight latent representations, and four training objectives, covering popular design choices in existing WAMs. We further validate our key findings on real-robot data from the DROID dataset. We hope to provide a systematic understanding of how core design choices affect world-action modeling and what principles can guide the development of future WAM systems.

Chao Tang, Haoqing Wang, Zilang Cen +9

Samsung Robotics eXperience · Samsung R&D Institute China–Beijing · Zhongguancun Academy +2

World Model Learning Causal World Models World Action Models Latent World Models World Model Evaluation

Sep 17, 2026cs.AI

Continual Enterprise World Model Discovery in Dynamic Systems

In an enterprise system, updating one field can set another, create a record, or start an approval. These effects are produced by business rules that are not built into the platform but written by each organization and revised over time. An agent working in such a system cannot predict the result of its own actions without knowing these rules. We study continual enterprise world model discovery, where an agent starts without knowledge of these business rules and discovers them by interacting with records and observing the outcomes. From those observations it builds a world model, which it revises as the rules change. To evaluate this, we introduce EnterpriseWorldShift, built on a live ServiceNow environment with nine tables, 25 hidden rules and 600 evaluation actions. It presents four versions of the same enterprise world, with the tables and records held fixed while a rule is modified, then added, then removed, so that discovery, revision, extension and retirement are each tested in turn. Our Continual Discovery Agent (CDA) builds such a model and carries it from one world to the next. It predicts the effects of the hidden rules more accurately than looking them up for each question, the approach taken by prior work, by up to 8.98 IoU points, and it answers from its own model without querying the running system.

Shambhavi Mishra, David Vazquez, Perouz Taslakian +3

ServiceNow Research · LIVIA, ILLS, ÉTS Montréal · Université Polytechnique Montréal +3

World Model Learning World Model Evaluation Interactive World Models

Sep 15, 2026cs.RO

World Models for Embodied Intelligence: From Plausible to Controllable to Actionable

World models connect perception and decision-making in embodied intelligence by maintaining hidden state, anticipating consequences, comparing interventions, and adapting when execution departs from expectations. Although progress is often measured by visual fidelity, their value lies in improving behavior. Before reaching for a cup, a person anticipates its weight and resistance to grasping, shaping the hand before contact. Such anticipation is coarse and rarely pictorial, yet it guides action. This raises a central question: which predictive capabilities improve behavior? Existing surveys, organized by architecture, output modality, or application domain, leave this question implicit. We introduce three progressively stronger capability levels: Plausible models preserve task-relevant temporal, geometric, or physical structure; Controllable models additionally predict how interventions alter that structure; and Actionable models translate predictions into measurable gains in planning, action, learning, evaluation, verification, recovery, or data selection. We complement this hierarchy with a 3 x 4 matrix crossing geometry, physics, and action grounding with improvement loops centered on data, rewards, policies, and the model itself. Using this framework, we survey manipulation, navigation, locomotion, autonomous driving, and general embodied learning, tracing technical progressions, clarifying capability requirements, and examining datasets, benchmarks, and evaluation protocols. We identify challenges in long-horizon consistency, uncertainty calibration, causal intervention testing, latency, verification and recovery, and cross-embodiment transfer. This perspective shifts evaluation from visual plausibility toward whether predictions capture task-relevant state, reflect intervention effects, and improve the closed-loop behavior of embodied agents.

Nanjie Yao, Hao Wang, Chong Cheng +10

HKUST(GZ) · University of Oxford · Princeton University +8

World Models Action-Conditioned World Models World Model Evaluation Embodied World Models

Sep 14, 2026cs.RO

IMPLY: Physically Anchored Consistency for World-Model Rollouts

A world model asked what happens if an object is pushed at several speeds produces several futures. If the model has the object in mind, those futures agree about it: each implies the same mass and friction. The consistency checks now used to vet world-action models ask whether a model's futures agree with each other, and none of them knows any physics. We show that this is not enough, and what to do instead. IMPLY reads the physics each rollout implies by inverting a simulator and scores a set of rollouts by how well one object explains all of them, anchored to two calibration pushes the model has observed. In a controlled setting, self-consistency gives a perfect score to a model that ignores the object and always predicts a typical push; anchoring exposes it (AUROC 0.70 versus 1.00). On a real model, V-JEPA 2-AC adapted to the scene, the same thing happens. Given its own calibration pushes the model tracks the object (per-object correlation with the truth 0.91); given another object's, it does not (0.05). Self-consistency cannot tell these apart, preferring the right evidence on 52% of objects, chance level, while anchored disagreement prefers it on 73% and correlates 0.92-0.99 with the rollouts' error. Used to choose among candidate rollout sets, it comes within 0.003 of an oracle that sees the truth. A model that has internalised the wrong object is exactly as self-consistent as one that has internalised the right one; consistency has to be anchored to evidence.

Aman Mehta, Riya Baviskar

Independent Researcher

World Models Action-Conditioned World Models World Model Evaluation

Sep 13, 2026cs.AI

One Model, Two Physical Stories: Auditing Misalignment in Multi-Modal World Modeling

World models, systems that generate what happens next given current environmental conditions, are increasingly being implemented with multi-modal generation in mind. However, generating multiple modalities simultaneously, such as visual simulations alongside physical state predictions in the form of text, introduces the risk of cross-modal inconsistency. Tested separately, both outputs may look convincing while still disagreeing: a model can calculate that a ball should rebound in one modality, then generate no rebound in another modality, to say nothing of diverging from real-world dynamics entirely. In this work we focus on two failures explicitly: \emph{Internal misalignment}, the disagreement between the world model's generated video and the same world model's prediction in a different modalities, and \emph{external misalignment} the disagreement between the world model's generation and an analytic physical environment. We derive common contracts of event, magnitude, timing, and construct a physics grounded pipeline to make comparisons measurable in both external and internal settings. We then ask whether progressively supplying the model's own contract (the A ladder for the internal setting) or a corrected physical contract (the B ladder for the external setting) closes the respective gaps. Across four mechanisms and 20 settings, we find that while language answers all 22 text probes correctly with respect to the true environment, the neutral video is often in disagreement, suggesting that the current unified backbones may not be capable of correct reasoning, internal consistency, and external physical fidelity all at once.

Geigh Zollicoffer, Minh Vu, Rajiv Ranasinghe +1

1Theoretical Division, Los Alamos National Laboratory · 2Computing and Artificial Intelligence Division, Los Alamos National Laboratory · 3Earth and Environmental Sciences Division, Los Alamos National Laboratory

Physical Consistency in Video Generation World Model Learning Cross-Modal Alignment World Model Evaluation Video World Models

Sep 13, 2026cs.LG

World Models for Cross-Machine CNC Transfer under Partial Sensor Overlap

Industrial world models must move between machines whose dynamics, sensing interfaces and command conventions differ. This study asks whether a command-conditioned latent world model, trained to predict future representations of the process rather than to reconstruct future samples, keeps its value on a machine it has never seen: a source CNC machine exposes 17 sensor channels, the target sharing 10 of those. All model selection uses source data only, and the locked configuration is evaluated on the target once. Two findings follow. First, latent-predictive pretraining brings no in-domain forecasting gain over matched training from scratch, so source accuracy alone cannot show what such a representation is worth. Second, the transferred model beats persistence on the unseen machine (with

R^2\approx0.01

against the target mean) but trails official forecasters that normalize each input window by its own statistics; a post-lock ablation, declared before it ran, shows that this input normalization alone closes the gap, and closing it costs predictive calibration. Cross-machine transfer under partial sensor overlap is therefore a distinct evaluation axis for command-conditioned world models.

Ayoub Louaye Bouaziz, Matthieu Ostertag, Anton Demasles

Université de Bretagne Occidentale, Brest, France · Mines Nancy, Université de Lorraine, Nancy, France

Manufacturing World Model Learning Action-Conditioned World Models Transfer Learning Predictive Representation Learning World Model Evaluation