mlsanityData to 14 Oct 2026

FantasyWorld: Geometry-Consistent World Modeling via Unified Video and 3D Prediction

arXiv 2509.21657 ↗First posted 25 Sep 2025cs.CVcited by 38

On the map: World models → Video-generation world models and simulators

Abstract (arXiv)

High-quality 3D world models are pivotal for embodied intelligence and Artificial General Intelligence (AGI), underpinning applications such as AR/VR content creation and robotic navigation. Despite the established strong imaginative priors, current video foundation models lack explicit 3D grounding capabilities, thus being limited in both spatial consistency and their utility for downstream 3D reasoning tasks. In this work, we present FantasyWorld, a geometry-enhanced framework that augments frozen video foundation models with a trainable geometric branch, enabling joint modeling of video latents and an implicit 3D field in a single forward pass. Our approach introduces cross-branch supervision, where geometry cues guide video generation and video priors regularize 3D prediction, thus yielding consistent and generalizable 3D-aware video representations. Notably, the resulting latents from the geometric branch can potentially serve as versatile representations for downstream 3D tasks such as novel view synthesis and navigation, without requiring per-scene optimization or fine-tuning. Extensive experiments show that FantasyWorld effectively bridges video imagination and 3D perception, outperforming recent geometry-consistent baselines in multi-view coherence and style consistency. Ablation studies further confirm that these gains stem from the unified backbone and cross-branch information exchange.

Key facts. Cited by 38 papers in the map (data through 14 Oct 2026). Placed in Video-generation world models and simulators, one of 11 directions in World models. Builds on 52 of its 55 references that are in the map. First posted on arXiv 25 Sep 2025.

Builds on

52 of 55 references.

Closest ancestors · 5

Same direction, newest first

  1. 2025Geometry Forcing: Marrying Video Diffusion and 3D Representation for Consistent World Modeling10 Jul
  2. 2025DeepVerse: 4D Autoregressive Video Generation as a World Model1 Jun
  3. 2025Wan: Open and Advanced Large-Scale Video Generative Models26 Mar
  4. 2022DreamFusion: Text-to-3D using 2D Diffusion29 Sep
  5. 2021Learning Transferable Visual Models From Natural Language Supervision26 Feb

Other references · 47

  1. 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
  2. 2023Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detectioncited by 3,332
  3. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
  4. 2023DUSt3R: Geometric 3D Vision Made Easycited by 1,567
  5. 2025VGGT: Visual Geometry Grounded Transformercited by 1,539
  6. 2023Fantasia3D: Disentangling Geometry and Appearance for High-quality Text-to-3D Content Creationcited by 687
  7. 2025Continuous 3D Perception Model with Persistent Statecited by 548
  8. 2023DL3DV-10K: A Large-Scale Scene Dataset for Deep Learning-based 3D Visioncited by 537
  9. 2020TartanAir: A Dataset to Push the Limits of Visual SLAMcited by 506
  10. 2024MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motioncited by 442
  11. 2024ViewCrafter: Taming Video Diffusion Models for High-fidelity Novel View Synthesiscited by 403
  12. 2025Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Passcited by 367
35 more
  1. 2025GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Controlcited by 341
  2. 2023Repurposing Diffusion-Based Image Generators for Monocular Depth Estimationcited by 327
  3. 2024GeoWizard: Unleashing the Diffusion Priors for 3D Geometry Estimation from a Single Imagecited by 286
  4. 20243D Reconstruction with Spatial Memorycited by 277
  5. 2020Infinite Nature: Perpetual View Generation of Natural Scenes from a Single Imagecited by 243
  6. 2025AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Viewscited by 243
  7. 2025Video Depth Anything: Consistent Depth Estimation for Super-Long Videoscited by 207
  8. 2024DepthCrafter: Generating Consistent Long Depth Sequences for Open-world Videoscited by 204
  9. 2024WonderWorld: Interactive 3D Scene Generation from a Single Imagecited by 202
  10. 20234D-fy: Text-to-4D Generation Using Hybrid Score Distillation Samplingcited by 185
  11. 2025FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Viewscited by 180
  12. 2025Context as Memory: Scene-Consistent Interactive Long Video Generation with Memory Retrievalcited by 163
  13. 2024MASt3R-SfM: a Fully-Integrated Solution for Unconstrained Structure-from-Motioncited by 162
  14. 2023WonderJourney: Going from Anywhere to Everywherecited by 158
  15. 2025WorldScore: A Unified Evaluation Benchmark for World Generationcited by 156
  16. 2023UniMax: Fairer and more Effective Language Sampling for Large-Scale Multilingual Pretrainingcited by 151
  17. 2024ReconX: Reconstruct Any Scene from Sparse Views with Video Diffusion Modelcited by 143
  18. 2024MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Secondscited by 141
  19. 2024DimensionX: Create Any 3D and 4D Scenes from a Single Image with Controllable Video Diffusioncited by 140
  20. 2024RGBD Objects in the Wild: Scaling Real-World 3D Object Learning from RGB-D Videoscited by 115
  21. 2025Voyager: Long-Range and World-Consistent Video Diffusion for Explorable 3D Scene Generationcited by 101
  22. 2024Wonderland: Navigating 3D Scenes from a Single Imagecited by 99
  23. 2025JavisDiT: Joint Audio-Video Diffusion Transformer with Hierarchical Spatio-Temporal Prior Synchronizationcited by 88
  24. 2025Uni3C: Unifying Precisely 3D-Enhanced Camera and Human Motion Controls for Video Generationcited by 77
  25. 2023UniPAD: A Universal Pre-training Paradigm for Autonomous Drivingcited by 73
  26. 20253DRS: MLLMs Need 3D-Aware Representation Supervision for Scene Understandingcited by 70
  27. 2024GaussianWorld: Gaussian World Model for Streaming 3D Occupancy Predictioncited by 67
  28. 2024World-consistent Video Diffusion with Explicit 3D Modelingcited by 65
  29. 2023PonderV2: Pave the Way for 3D Foundation Model with A Universal Pre-training Paradigmcited by 56
  30. 2025Matrix3D: Large Photogrammetry Model All-in-Onecited by 40
  31. 2024Vidu4D: Single Generated Video to High-Fidelity 4D Reconstruction with Dynamic Gaussian Surfelscited by 38
  32. 2025Matrix-3D: Omnidirectional Explorable 3D World Generationcited by 35
  33. 2024SPA: 3D Spatial-Awareness Enables Effective Embodied Representationcited by 28
  34. 2025WorldExplorer: Towards Generating Fully Navigable 3D Scenescited by 26
  35. 2024On the Feature Learning in Diffusion Modelscited by 20

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 3; 2026: 35.

38 papers.

Sort:

By area:

  1. 2026Lyra 2.0: Explorable Generative 3D Worldscited by 44
  2. 2026WBench: A Comprehensive Multi-turn Benchmark for Interactive Video World Model Evaluationcited by 39
  3. 2026VGGRPO: Towards World-Consistent Video Generation with 4D Latent Rewardcited by 24
  4. 2026INSPATIO-WORLD: A Real-Time 4D World Simulator via Spatiotemporal Autoregressive Modelingcited by 21
  5. 2026Omni-WorldBench: Towards a Comprehensive Interaction-Centric Evaluation for World Modelscited by 18
  6. 2026Grounding World Simulation Models in a Real-World Metropoliscited by 12
  7. 2026MemoBench: Benchmarking World Modeling in Dynamically Changing Environmentscited by 12
  8. 2026World-R1: Reinforcing 3D Constraints for Text-to-Video Generationcited by 11
  9. 2026Towards Interactive Video World Modeling: Frontiers, Challenges, Benchmarks, and Future Trendscited by 8
  10. 2026CineScene: Implicit 3D as Effective Scene Representation for Cinematic Video Generationcited by 7
  11. 2026VIGOR: VIdeo Geometry-Oriented Reward for Temporal Generative Alignmentcited by 7
  12. 2026VideoWeave: Unlocking Geometric Consistency in Video Generation via Joint Geometry-Video Modelingcited by 7
26 more
  1. 2026Feed-Forward 3D Scene Modeling: A Problem-Driven Perspectivecited by 6
  2. 2025WorldReel: 4D Video Generation with Consistent Geometry and Motion Modelingcited by 5
  3. 2026Warp-as-History: Generalizable Camera-Controlled Video Generation from One Training Videocited by 4
  4. 2026PixWorld: Unifying 3D Scene Generation and Reconstruction in Pixel Spacecited by 4
  5. 2026WorldExam: Benchmarking World Models from Apparent Appearance to Inherent Reactivitycited by 4
  6. 2026HarnessEval-W: Agentifying the Evaluation of Visual Worldscited by 4
  7. 2025MagicWorld: Towards Long-Horizon Stability for Interactive Video World Explorationcited by 3
  8. 2026OneWorld: Taming Scene Generation with 3D Unified Representation Autoencodercited by 2
  9. 2026GTA: Advancing Image-to-3D World Generation via Geometry Then Appearance Video Diffusioncited by 2
  10. 2026Drift-Resistant Navigation World Model with Anchored Epipolar Guidancecited by 2
  11. 2026Envision4D: Envisioning Visual Futures via Feed-forward 4D Gaussian Splatting for Autonomous Drivingcited by 1
  12. 2026OctWorld: Long-Range World-Consistent Video Generation with Octree-Based 3D Mappingcited by 1
  13. 2026Precise Editing and Flexible Referencing for Interactable Worldscited by 1
  14. 2025Envision: Embodied Visual Planning via Goal-Imagery Video Diffusioncited by 0
  15. 2026SceneExpander: Text-Guided 3D Scene Expansion via Free-Form View Insertioncited by 0
  16. 2026MMDiff: Extending Diffusion Transformers for Multi-Modal Generationcited by 0
  17. 2026RoGe: Novel View Synthesis via End-to-End Implicit Reconstruction and Generationcited by 0
  18. 2026Stabilizing Camera-Controlled Novel View Synthesis at Inference Timecited by 0
  19. 2026Puffin-World: Scaling a Unified Multimodal Model with Native 3D World Statescited by 0
  20. 2026World in World: Explore the World with World Modelscited by 0
  21. 2026Proxy2World: Learning to Generate Worlds From Lightweight Proxies without Seeing Themcited by 0
  22. 2026LongLive-Plug: Once-for-All Distillation for Video Generationcited by 0
  23. 2026World Observer: Joint Actor-Observer Generation for Persistent World Modelingcited by 0
  24. 2026WorldAlign: Decoupled 4D Reward for World-Consistent Video Generationcited by 0
  25. 2026GAD: Geometry Attention Distillation for Consistent Video Generationcited by 0
  26. 2026Any Time, Any Space: A World-State-Centric Framework for Agentic Video Generationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025From 2D to 3D Cognition: A Brief Survey of General World Models
  2. 2026PanoWorld: Geometry-Consistent Panoramic Video World Modeling
  3. 20263D-Belief: Embodied Belief Inference via Generative 3D World Modeling
  4. 2026DreamWorld: Unified World Modeling in Video Generation
  5. 2026DriveDreamer-Policy: A Geometry-Grounded World-Action Model for Unified Generation and Planning
  6. 2026Quantitative Video World Model Evaluation for Geometric-Consistency
  7. 20254DWorldBench: A Comprehensive Evaluation Framework for 3D/4D World Generation Models
  8. 2026Geometry-Aware Implicit Memory for Video World Models