VLA-Arena: An Open-Source Framework for Benchmarking Vision-Language-Action Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Borong, Li, Jiahao, Shen, Jiachen, Cai, Yishuai, Zhang, Yuhao, Chen, Yuanpei, Dai, Juntao, Ji, Jiaming, Yang, Yaodong
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909976822284288
author Zhang, Borong
Li, Jiahao
Shen, Jiachen
Cai, Yishuai
Zhang, Yuhao
Chen, Yuanpei
Dai, Juntao
Ji, Jiaming
Yang, Yaodong
author_facet Zhang, Borong
Li, Jiahao
Shen, Jiachen
Cai, Yishuai
Zhang, Yuhao
Chen, Yuanpei
Dai, Juntao
Ji, Jiaming
Yang, Yaodong
contents While Vision-Language-Action models (VLAs) are rapidly advancing towards generalist robot policies, it remains difficult to quantitatively understand their limits and failure modes. To address this, we introduce a comprehensive benchmark called VLA-Arena. We propose a novel structured task design framework to quantify difficulty across three orthogonal axes: (1) Task Structure, (2) Language Command, and (3) Visual Observation. This allows us to systematically design tasks with fine-grained difficulty levels, enabling a precise measurement of model capability frontiers. For Task Structure, VLA-Arena's 170 tasks are grouped into four dimensions: Safety, Distractor, Extrapolation, and Long Horizon. Each task is designed with three difficulty levels (L0-L2), with fine-tuning performed exclusively on L0 to assess general capability. Orthogonal to this, language (W0-W4) and visual (V0-V4) perturbations can be applied to any task to enable a decoupled analysis of robustness. Our extensive evaluation of state-of-the-art VLAs reveals several critical limitations, including a strong tendency toward memorization over generalization, asymmetric robustness, a lack of consideration for safety constraints, and an inability to compose learned skills for long-horizon tasks. To foster research addressing these challenges and ensure reproducibility, we provide the complete VLA-Arena framework, including an end-to-end toolchain from task definition to automated evaluation and the VLA-Arena-S/M/L datasets for fine-tuning. Our benchmark, data, models, and leaderboard are available at https://vla-arena.github.io.
format Preprint
id arxiv_https___arxiv_org_abs_2512_22539
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle VLA-Arena: An Open-Source Framework for Benchmarking Vision-Language-Action Models
Zhang, Borong
Li, Jiahao
Shen, Jiachen
Cai, Yishuai
Zhang, Yuhao
Chen, Yuanpei
Dai, Juntao
Ji, Jiaming
Yang, Yaodong
Robotics
Computer Vision and Pattern Recognition
While Vision-Language-Action models (VLAs) are rapidly advancing towards generalist robot policies, it remains difficult to quantitatively understand their limits and failure modes. To address this, we introduce a comprehensive benchmark called VLA-Arena. We propose a novel structured task design framework to quantify difficulty across three orthogonal axes: (1) Task Structure, (2) Language Command, and (3) Visual Observation. This allows us to systematically design tasks with fine-grained difficulty levels, enabling a precise measurement of model capability frontiers. For Task Structure, VLA-Arena's 170 tasks are grouped into four dimensions: Safety, Distractor, Extrapolation, and Long Horizon. Each task is designed with three difficulty levels (L0-L2), with fine-tuning performed exclusively on L0 to assess general capability. Orthogonal to this, language (W0-W4) and visual (V0-V4) perturbations can be applied to any task to enable a decoupled analysis of robustness. Our extensive evaluation of state-of-the-art VLAs reveals several critical limitations, including a strong tendency toward memorization over generalization, asymmetric robustness, a lack of consideration for safety constraints, and an inability to compose learned skills for long-horizon tasks. To foster research addressing these challenges and ensure reproducibility, we provide the complete VLA-Arena framework, including an end-to-end toolchain from task definition to automated evaluation and the VLA-Arena-S/M/L datasets for fine-tuning. Our benchmark, data, models, and leaderboard are available at https://vla-arena.github.io.
title VLA-Arena: An Open-Source Framework for Benchmarking Vision-Language-Action Models
topic Robotics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.22539