What Matters in Quality-Aware VLA Pretraining: A Benchmark and Empirical Study

arXiv:2610.18869 · cs.RO, cs.AI, cs.LG · Submitted 2026-10-13 · Read on arXiv

cs.RO, cs.AI, cs.LG

Submitted: 2026-10-13

Updated: 2026-10-13

Comments: Project Website: https://axis-bench.github.io/

Project page: https://axis-bench.github.io

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Benchmark — 1× in this paper · explained in 2 episodes

Sources

• ${\pi}_{0.7}$: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities
• $\pi^{*}_{0.6}$: a VLA That Learns From Experience
• World Value Models for Robotic Manipulation
• Advantage-Weighted Regression: Simple and Scalable Off-Policy Reinforcement Learning
• SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics
• AXIS: A Growable Community-Driven Data Engine for Scalable Robot Manipulation
• robosuite: A Modular Simulation Framework and Benchmark for Robot Learning

Related papers

• FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
• MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
• RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
• ARC: A Reasoning Recipe for Robot Foundation Models
• HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
• APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies