# One Token Per Frame: Reconsidering Visual Bandwidth in World Models for VLA Policy

arXiv: https://arxiv.org/abs/2605.07931 · HTML: https://mlsanity.com/paper/2605.07931/

## Key facts

- Cited by 6 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 54 of its 60 references that are in the map.
- First posted on arXiv 8 May 2026.

## Placed in

- World models / [Embodied and robot world models](https://mlsanity.com/wm/embodied-and-robot-world-models/)
- Vision-language-action models / [VLA architectures and generalization](https://mlsanity.com/vla/vla-architectures-and-generalization/)

## Builds on (most cited first, 30 of 54)

- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [Diffusion Policy: Visuomotor Policy Learning via Action Diffusion](https://mlsanity.com/paper/2303.04137/) (2023)
- [RT-1: Robotics Transformer for Real-World Control at Scale](https://mlsanity.com/paper/2212.06817/) (2022)
- [π0.5: a Vision-Language-Action Model with Open-World Generalization](https://mlsanity.com/paper/2504.16054/) (2025)
- [Dream to Control: Learning Behaviors by Latent Imagination](https://mlsanity.com/paper/1912.01603/) (2019)
- [Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware](https://mlsanity.com/paper/2304.13705/) (2023)
- [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://mlsanity.com/paper/2106.01345/) (2021)
- [Octo: An Open-Source Generalist Robot Policy](https://mlsanity.com/paper/2405.12213/) (2024)
- [LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning](https://mlsanity.com/paper/2306.03310/) (2023)
- [GR00T N1: An Open Foundation Model for Generalist Humanoid Robots](https://mlsanity.com/paper/2503.14734/) (2025)
- [Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture](https://mlsanity.com/paper/2301.08243/) (2023)
- [Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success](https://mlsanity.com/paper/2502.19645/) (2025)
- [PaliGemma: A versatile 3B VLM for transfer](https://mlsanity.com/paper/2407.07726/) (2024)
- [R3M: A Universal Visual Representation for Robot Manipulation](https://mlsanity.com/paper/2203.12601/) (2022)
- [RDT-1B: a Diffusion Foundation Model for Bimanual Manipulation](https://mlsanity.com/paper/2410.07864/) (2024)
- [Revisiting Feature Prediction for Learning Visual Representations from Video](https://mlsanity.com/paper/2404.08471/) (2024)
- [SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics](https://mlsanity.com/paper/2506.01844/) (2025)
- [Learning Universal Policies via Text-Guided Video Generation](https://mlsanity.com/paper/2302.00111/) (2023)
- [SpatialVLA: Exploring Spatial Representations for Visual-Language-Action Model](https://mlsanity.com/paper/2501.15830/) (2025)
- [CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models](https://mlsanity.com/paper/2503.22020/) (2025)
- [CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulation](https://mlsanity.com/paper/2411.19650/) (2024)
- [UniVLA: Learning to Act Anywhere with Task-centric Latent Actions](https://mlsanity.com/paper/2505.06111/) (2025)
- [VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training](https://mlsanity.com/paper/2210.00030/) (2022)
- [WorldVLA: Towards Autoregressive Action World Model](https://mlsanity.com/paper/2506.21539/) (2025)
- [Vision-Language Foundation Models as Effective Robot Imitators](https://mlsanity.com/paper/2311.01378/) (2023)
- [3D-VLA: A 3D Vision-Language-Action Generative World Model](https://mlsanity.com/paper/2403.09631/) (2024)

## Built on by (most cited first, 6 of 6)

- [ABot-M0.5: Unified Mobility-and-Manipulation World Action Model](https://mlsanity.com/paper/2607.00678/) (2026)
- [ALAM: Algebraically Consistent Latent Action Model for Vision-Language-Action Models](https://mlsanity.com/paper/2605.10819/) (2026)
- [World Tokens: Enhancing Embodied Policies with Training-Time World Modeling](https://mlsanity.com/paper/2608.09730/) (2026)
- [DLAM: Distributional Latent Actions with Temporal Constraints](https://mlsanity.com/paper/2607.27138/) (2026)
- [SLIM-0.5B: Learning Action-Grounded Predictive Latents for Robot Manipulation](https://mlsanity.com/paper/2608.09771/) (2026)
- [Token-World: World Modeling in Vision-Language Model Token Space for Robot Manipulation](https://mlsanity.com/paper/2610.00575/) (2026)

## Abstract (arXiv)

Vision-language-action (VLA) models can use visual prediction to anticipate future states, but dense visual features make the generative sequence grow with the number of camera views, prediction horizon, and encoder resolution. Whether such dense representations are necessary for effective control remains unclear. We introduce OneWM-VLA, which represents each retained camera view with one predictive token per future step. Adaptive Attention Pooling compresses visual features into compact latents, which are jointly generated with robot actions under a conditional flow-matching objective. Future observations provide the latent targets during training and are not required at inference. This design incorporates visual prediction into a pretrained VLA policy while keeping the generative sequence compact. On MetaWorld~MT50, OneWM-VLA improves the average success rate of the π0 backbone from 47.91% to 61.53%, reaching 72.01% after 60k training steps. It also achieves 98.1% success on LIBERO and raises Fold Cloth success on a real Piper arm from 20.0% to 60.0% relative to π0. Comparisons on two additional VLA backbones consistently favor one token over three across the evaluated checkpoints. A matched ablation at a longer action horizon further shows that removing the latent loss reduces success from 58.09% to 21.64%, supporting the benefit of future supervision for policy learning.