# AnyStep-WAM: Budget-Aligned Distillation and Adaptive Inference for World Action Models

arXiv: https://arxiv.org/abs/2609.33748 · HTML: https://mlsanity.com/paper/2609.33748/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Embodied and robot world models, one of 11 directions in World models.
- Builds on 19 of its 28 references that are in the map.
- First posted on arXiv 27 Sep 2026.

## Placed in

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

## Builds on (most cited first, 19 of 19)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference](https://mlsanity.com/paper/2310.04378/) (2023)
- [RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation](https://mlsanity.com/paper/2506.18088/) (2025)
- [Fast-WAM: Do World Action Models Need Test-time Future Imagination?](https://mlsanity.com/paper/2603.16666/) (2026)
- [Causal World Modeling for Robot Control](https://mlsanity.com/paper/2601.21998/) (2026)
- [Improved Mean Flows: On the Challenges of Fastforward Generative Models](https://mlsanity.com/paper/2512.02012/) (2025)
- [ManiFlow: A General Robot Manipulation Policy via Consistency Flow Training](https://mlsanity.com/paper/2509.01819/) (2025)
- [Flash-WAM: Modality-Aware Distillation for World Action Models](https://mlsanity.com/paper/2606.05254/) (2026)
- [Stable Velocity: A Variance Perspective on Flow Matching](https://mlsanity.com/paper/2602.05435/) (2026)
- [D3P: Dynamic Denoising Diffusion Policy via Reinforcement Learning](https://mlsanity.com/paper/2508.06804/) (2025)
- [DSA: Dynamic Step Allocation for Fast Autoregressive Video Generation](https://mlsanity.com/paper/2606.04432/) (2026)
- [Let It Be Simple: One-Step Action Generation for Vision-Language-Action Models](https://mlsanity.com/paper/2606.05737/) (2026)
- [ELASTIC: Efficiently Learning to Adaptively Scale Test-Time Compute for Generative Control Policies](https://mlsanity.com/paper/2606.31132/) (2026)
- [The Geometry of Flow-Matching Uncertainty: A Cost-free Uncertainty Proxy and Its Application in Flow-based VLA Failure Detection](https://mlsanity.com/paper/2607.27933/) (2026)
- [DriftingVLA: Native One-Step Vision-Language-Action Generation via Per-Dimension Temporal Drifting](https://mlsanity.com/paper/2608.29749/) (2026)
- [Elastic Queries Reinforcement Learning: Self-Aware Policy Execution for VLA Models](https://mlsanity.com/paper/2606.14375/) (2026)
- [Adaptive-WAM: Quality-Guided Early-Exit Planning from Intermediate Video-Diffusion Features](https://mlsanity.com/paper/2608.06008/) (2026)
- [AdaVLA: Adaptive Step Flow Matching for Training-free Acceleration of Vision-Language-Action Models](https://mlsanity.com/paper/2608.29208/) (2026)

## Abstract (arXiv)

World-action models (WAMs) couple predictive visual modeling with action generation, typically relying on iterative denoising with a fixed denoising steps. However, manipulation tasks contain actions chunks with varying sensitivity to generation errors: critical actions require precision, while less sensitive actions allow faster generation with fewer denoising steps. Here we introduce AnyStep World Action Model, a general framework for tunable-budget prediction and scene-dependent computation allocation. Our budget-aligned teacher-trajectory distillation trains interval-conditioned flow maps using explicit frozen-teacher transitions and shared low-rank adapters, supporting action generation from one-step prediction to multi-step refinement. Building on this capability, a lightweight risk-benefit scheduler predicts teacher-curvature-based difficulty and budget-specific student-teacher fidelity from a single one-step preview, selecting the smallest budget predicted to satisfy risk-adaptive fidelity requirements. We evaluate our framework on three widely used WAMs Motus, FastWAM, and LingBotVA using RoboTwin 2.0. Our method reduces average denoising steps by 60.2%, 49.8%, and 85.28%, respectively, while maintaining baseline task success rates. In particular, our AnyStep training substantially improves model performance under a one-step denoising budget, increasing task success rates by 7.07%, 12.08%, and 8.94% on Motus, FastWAM, and LingBotVA, respectively. Experiments on six real-world manipulation tasks further validate its effectiveness.