# KeyWorld: Key Frame Reasoning Enables Effective and Efficient World Models

arXiv: https://arxiv.org/abs/2509.21027 · HTML: https://mlsanity.com/paper/2509.21027/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Placed in Embodied and robot world models, one of 11 directions in World models.
- Builds on 18 of its 24 references that are in the map.
- First posted on arXiv 25 Sep 2025.

## Placed in

- World models / [Embodied and robot world models](https://mlsanity.com/wm/embodied-and-robot-world-models/)

## Builds on (most cited first, 18 of 18)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer](https://mlsanity.com/paper/2408.06072/) (2024)
- [RT-1: Robotics Transformer for Real-World Control at Scale](https://mlsanity.com/paper/2212.06817/) (2022)
- [LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning](https://mlsanity.com/paper/2306.03310/) (2023)
- [Cosmos World Foundation Model Platform for Physical AI](https://mlsanity.com/paper/2501.03575/) (2025)
- [TD-MPC2: Scalable, Robust World Models for Continuous Control](https://mlsanity.com/paper/2310.16828/) (2023)
- [WorldVLA: Towards Autoregressive Action World Model](https://mlsanity.com/paper/2506.21539/) (2025)
- [Unified Video Action Model](https://mlsanity.com/paper/2503.00200/) (2025)
- [RoboDreamer: Learning Compositional World Models for Robot Imagination](https://mlsanity.com/paper/2404.12377/) (2024)
- [Understanding World or Predicting Future? A Comprehensive Survey of World Models](https://mlsanity.com/paper/2411.14499/) (2024)
- [FILM: Frame Interpolation for Large Motion](https://mlsanity.com/paper/2202.04901/) (2022)
- [iVideoGPT: Interactive VideoGPTs are Scalable World Models](https://mlsanity.com/paper/2405.15223/) (2024)
- [WorldDreamer: Towards General World Models for Video Generation via Predicting Masked Tokens](https://mlsanity.com/paper/2401.09985/) (2024)
- [Real-World Robot Applications of Foundation Models: A Review](https://mlsanity.com/paper/2402.05741/) (2024)
- [RoboScape: Physics-informed Embodied World Model](https://mlsanity.com/paper/2506.23135/) (2025)
- [RoboEnvision: A Long-Horizon Video Generation Model for Multi-Task Robot Manipulation](https://mlsanity.com/paper/2506.22007/) (2025)
- [Reward-Consistent Dynamics Models are Strongly Generalizable for Offline Reinforcement Learning](https://mlsanity.com/paper/2310.05422/) (2023)
- [FusionFrames: Efficient Architectural Aspects for Text-to-Video Generation Pipeline](https://mlsanity.com/paper/2311.13073/) (2023)

## Built on by (most cited first, 7 of 7)

- [Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application](https://mlsanity.com/paper/2606.12191/) (2026)
- [SKIP: Sparse Keyframe Interpolation Paradigm for Efficient Embodied World Models](https://mlsanity.com/paper/2606.00664/) (2026)
- [DVG-WM: Disentangled Video Generation Enables Efficient Embodied World Model for Robotic Manipulation](https://mlsanity.com/paper/2606.32028/) (2026)
- [Semantic Communications with World Models](https://mlsanity.com/paper/2510.24785/) (2025)
- [Improving Robotic Imitation Learning via Trajectory Standardization](https://mlsanity.com/paper/2606.22907/) (2026)
- [Toward Unified Robot Learning: Bridging Representation, Vision-Language-Action, and World Models](https://mlsanity.com/paper/2609.03927/) (2026)
- [STRIKE: Learning Visual State Transitions for Physical World Modeling](https://mlsanity.com/paper/2610.09514/) (2026)

## Abstract (arXiv)

Robotic world models are a promising paradigm for forecasting future environment states, yet their inference speed and the physical plausibility of generated trajectories remain critical bottlenecks, limiting their real-world applications. This stems from the redundancy of the prevailing frame-to-frame generation approach, where the model conducts costly computation on similar frames, as well as neglecting the semantic importance of key transitions. To address this inefficiency, we propose KeyWorld, a framework that improves text-conditioned robotic world models by concentrating transformers computation on a few semantic key frames while employing a lightweight convolutional model to fill the intermediate frames. Specifically, KeyWorld first identifies significant transitions by iteratively simplifying the robot's motion trajectories, obtaining the ground truth key frames. Then, a DiT model is trained to reason and generate these physically meaningful key frames from textual task descriptions. Finally, a lightweight interpolator efficiently reconstructs the full video by inpainting all intermediate frames. Evaluations on the LIBERO benchmark demonstrate that KeyWorld achieves a 5.68× acceleration compared to the frame-to-frame generation baseline, and focusing on the motion-aware key frames further contributes to the physical validity of the generated videos, especially on complex tasks. Our approach highlights a practical path toward deploying world models in real-time robotic control and other domains requiring both efficient and effective world models. Code is released at https://anonymous.4open.science/r/Keyworld-E43D.