# ELASTIC: Efficiently Learning to Adaptively Scale Test-Time Compute for Generative Control Policies

arXiv: https://arxiv.org/abs/2606.31132 · HTML: https://mlsanity.com/paper/2606.31132/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 8 of its 36 references that are in the map.
- First posted on arXiv 30 Jun 2026.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)
- Diffusion models / [Diffusion and flow policies for robotics](https://mlsanity.com/diffusion/diffusion-and-flow-policies-for-robotics/)

## Builds on (most cited first, 8 of 8)

- [RoboMonkey: Scaling Test-Time Sampling and Verification for Vision-Language-Action Models](https://mlsanity.com/paper/2506.17811/) (2025)
- [RoboReward: General-Purpose Vision-Language Reward Models for Robotics](https://mlsanity.com/paper/2601.00675/) (2026)
- [Robo-Dopamine: General Process Reward Modeling for High-Precision Robotic Manipulation](https://mlsanity.com/paper/2512.23703/) (2025)
- [Learning to Reason Across Parallel Samples for LLM Reasoning](https://mlsanity.com/paper/2506.09014/) (2025)
- [D3P: Dynamic Denoising Diffusion Policy via Reinforcement Learning](https://mlsanity.com/paper/2508.06804/) (2025)
- [Zero-Overhead Introspection for Adaptive Test-Time Compute](https://mlsanity.com/paper/2512.01457/) (2025)
- [FASTER: Value-Guided Sampling for Fast RL](https://mlsanity.com/paper/2604.19730/) (2026)
- [Adaptive Time Step Flow Matching for Autonomous Driving Motion Planning](https://mlsanity.com/paper/2602.10285/) (2026)

## Built on by (most cited first, 3 of 3)

- [Think Fast, Plan Selectively: Adaptive Deliberation for Efficient Data-Driven MPC](https://mlsanity.com/paper/2609.32591/) (2026)
- [AnyStep-WAM: Budget-Aligned Distillation and Adaptive Inference for World Action Models](https://mlsanity.com/paper/2609.33748/) (2026)
- [ReSync: Re-Aligning the Two Clocks of Asynchronous World-Action Models](https://mlsanity.com/paper/2609.33944/) (2026)

## Abstract (arXiv)

Generative control policies (GCPs), such as diffusion policies and flow-based vision-language-action models, enable test-time scaling in robot control. Test-time compute can be allocated along two axes: sequential scaling, which increases denoising steps to refine actions, and parallel scaling, which samples multiple candidate actions to search across modes of the policy distribution. However, the optimal allocation of sequential and parallel compute is hard to know a priori as it is state-, task-, and policy-dependent. For example, early stages of a grasp may benefit from broader parallel exploration, while near-contact phases may require more sequential refinement for precision. We present ELASTIC, an algorithm that learns state-dependent test-time compute schedules for GCPs. We formulate compute allocation as a meta-Markov Decision Process in which a meta-policy interacts with a frozen pretrained robot policy and selects sequential steps and parallel samples at each denoising iteration to maximize task success while minimizing compute. Using reinforcement learning, this meta-policy also learns adaptive compute schedules without access to the GCP's training data. Across simulated manipulation benchmarks with diffusion policies, ELASTIC Pareto-dominates fixed and single-axis scaling baselines at matched compute budgets. On real-world robot manipulation with the π0.5 vision-language-action model, ELASTIC matches best-of-10 success while reducing wall-clock latency by 34%.