# WorldCompass: Reinforcement Learning for Long-Horizon World Models

arXiv: https://arxiv.org/abs/2602.09022 · HTML: https://mlsanity.com/paper/2602.09022/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- Placed in Latent dynamics models and model-based RL, one of 11 directions in World models.
- Builds on 39 of its 44 references that are in the map.
- First posted on arXiv 9 Feb 2026.

## Placed in

- World models / [Latent dynamics models and model-based RL](https://mlsanity.com/wm/latent-dynamics-models-and-model-based-rl/)
- Deep reinforcement learning methods / [RL for robot locomotion and manipulation](https://mlsanity.com/deeprl/rl-for-robot-locomotion-and-manipulation/)

## Builds on (most cited first, 30 of 39)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Score-Based Generative Modeling through Stochastic Differential Equations](https://mlsanity.com/paper/2011.13456/) (2020)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [Cosmos World Foundation Model Platform for Physical AI](https://mlsanity.com/paper/2501.03575/) (2025)
- [Group Sequence Policy Optimization](https://mlsanity.com/paper/2507.18071/) (2025)
- [DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models](https://mlsanity.com/paper/2512.02556/) (2025)
- [Diffusion Model Alignment Using Direct Preference Optimization](https://mlsanity.com/paper/2311.12908/) (2023)
- [Genie: Generative Interactive Environments](https://mlsanity.com/paper/2402.15391/) (2024)
- [Depth Anything 3: Recovering the Visual Space from Any Views](https://mlsanity.com/paper/2511.10647/) (2025)
- [Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusion](https://mlsanity.com/paper/2407.01392/) (2024)
- [Flow-GRPO: Training Flow Matching Models via Online RL](https://mlsanity.com/paper/2505.05470/) (2025)
- [MotionCtrl: A Unified and Flexible Motion Controller for Video Generation](https://mlsanity.com/paper/2312.03641/) (2023)
- [Kimi K2: Open Agentic Intelligence](https://mlsanity.com/paper/2507.20534/) (2025)
- [CameraCtrl: Enabling Camera Control for Text-to-Video Generation](https://mlsanity.com/paper/2404.02101/) (2024)
- [DanceGRPO: Unleashing GRPO on Visual Generation](https://mlsanity.com/paper/2505.07818/) (2025)
- [Diffusion Models Are Real-Time Game Engines](https://mlsanity.com/paper/2408.14837/) (2024)
- [DiffusionNFT: Online Diffusion Reinforcement with Forward Process](https://mlsanity.com/paper/2509.16117/) (2025)
- [HPSv3: Towards Wide-Spectrum Human Preference Score](https://mlsanity.com/paper/2508.03789/) (2025)
- [World Simulation with Video Foundation Models for Physical AI](https://mlsanity.com/paper/2511.00062/) (2025)
- [Matrix-game 2.0: An open-source real-time and streaming interactive world model](https://mlsanity.com/paper/2508.13009/) (2025)
- [Aligning Cyber Space with Physical World: A Comprehensive Survey on Embodied AI](https://mlsanity.com/paper/2407.06886/) (2024)
- [WorldPlay: Towards Long-Term Geometric Consistency for Real-Time Interactive World Modeling](https://mlsanity.com/paper/2512.14614/) (2025)
- [Understanding World or Predicting Future? A Comprehensive Survey of World Models](https://mlsanity.com/paper/2411.14499/) (2024)
- [Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning](https://mlsanity.com/paper/2503.15558/) (2025)
- [HunyuanVideo 1.5 Technical Report](https://mlsanity.com/paper/2511.18870/) (2025)
- [Genie Envisioner: A Unified World Foundation Platform for Robotic Manipulation](https://mlsanity.com/paper/2508.05635/) (2025)
- [GameFactory: Creating New Games with Generative Interactive Videos](https://mlsanity.com/paper/2501.08325/) (2025)
- [Voyager: Long-Range and World-Consistent Video Diffusion for Explorable 3D Scene Generation](https://mlsanity.com/paper/2506.04225/) (2025)

## Built on by (most cited first, 24 of 24)

- [Matrix-Game 3.0: Real-Time and Streaming Interactive World Model with Long-Horizon Memory](https://mlsanity.com/paper/2604.08995/) (2026)
- [Causal Forcing++: Scalable Few-Step Autoregressive Diffusion Distillation for Real-Time Interactive Video Generation](https://mlsanity.com/paper/2605.15141/) (2026)
- [HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worlds](https://mlsanity.com/paper/2604.14268/) (2026)
- [DreamX-World 1.0: A General-Purpose Interactive World Model](https://mlsanity.com/paper/2606.16993/) (2026)
- [Reward Hacking in the Era of Large Models: Mechanisms, Emergent Misalignment, Challenges](https://mlsanity.com/paper/2604.13602/) (2026)
- [VGGRPO: Towards World-Consistent Video Generation with 4D Latent Reward](https://mlsanity.com/paper/2603.26599/) (2026)
- [AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories](https://mlsanity.com/paper/2602.14941/) (2026)
- [ActWorld: From Explorable to Interactive World Model via Action-Aware Memory](https://mlsanity.com/paper/2606.17730/) (2026)
- [Persistent Robot World Models: Stabilizing Multi-Step Rollouts via Reinforcement Learning](https://mlsanity.com/paper/2603.25685/) (2026)
- [WorldRoamBench: An Open-World Benchmark for Long-Horizon Stability of Interactive World Models](https://mlsanity.com/paper/2606.31672/) (2026)
- [Towards Interactive Video World Modeling: Frontiers, Challenges, Benchmarks, and Future Trends](https://mlsanity.com/paper/2606.01164/) (2026)
- [RAVEN: Real-time Autoregressive Video Extrapolation with Consistency-model GRPO](https://mlsanity.com/paper/2605.15190/) (2026)
- [PhyMotion: Structured 3D Motion Reward for Physics-Grounded Human Video Generation](https://mlsanity.com/paper/2605.14269/) (2026)
- [WorldCycle: Self-Verifiable Reinforcement Learning for Long-Horizon Video World Models](https://mlsanity.com/paper/2608.04964/) (2026)
- [CreFlow: Corrective Reflow for Sparse-Reward Embodied Video Diffusion RL](https://mlsanity.com/paper/2605.14274/) (2026)
- [OnlineWM: Causality-Aware Active Online Learning for Effective World Modeling](https://mlsanity.com/paper/2609.23753/) (2026)
- [PAVXploreRL: Physical-Action-Visual World Model Reinforcement Learning with Action Exploration](https://mlsanity.com/paper/2607.16602/) (2026)
- [Stream4D: 4D-Consistency for Streaming Autoregressive Diffusion Video Models](https://mlsanity.com/paper/2608.19556/) (2026)
- [WorldReward: Reward Modeling for Camera-Conditioned World Models](https://mlsanity.com/paper/2609.03952/) (2026)
- [WorldPlay2: Extending Real-Time Interactive World Models in Control and Horizon](https://mlsanity.com/paper/2609.35560/) (2026)
- [DynaTokens: Teaching Dynamics to Camera-Controlled Video Models at Test Time](https://mlsanity.com/paper/2609.35704/) (2026)
- [Spatial Memory Intelligence: Endowing World Models with Understanding-Driven Long-Term Memory](https://mlsanity.com/paper/2610.02521/) (2026)
- [LoGo: Local-Global Rewards for Consistent Long-Horizon Video Generation](https://mlsanity.com/paper/2610.03636/) (2026)
- [DreamTrue: Action-Faithful Robot World Model with Counterfactual Post-Training](https://mlsanity.com/paper/2610.12468/) (2026)

## Abstract (arXiv)

This work presents WorldCompass, a novel Reinforcement Learning (RL) post-training framework for the long-horizon, interactive video-based world models, enabling them to explore the world more accurately and consistently based on interaction signals. To effectively "steer" the world model's exploration, we introduce three core innovations tailored to the autoregressive video generation paradigm: 1) Clip-level rollout Strategy: We generate and evaluate multiple samples at a single target clip, which significantly boosts rollout efficiency and provides fine-grained reward signals. 2) Complementary Reward Functions: We design reward functions for both interaction-following accuracy and visual quality, which provide direct supervision and effectively suppress reward-hacking behaviors. 3) Efficient RL Algorithm: We employ the negative-aware fine-tuning strategy coupled with various efficiency optimizations to efficiently and effectively enhance model capacity. Evaluations on the SoTA open-source world model, WorldPlay, demonstrate that WorldCompass significantly improves interaction accuracy and visual fidelity across various scenarios.