aiqbench analysis
AnolisVLA introduces a lightweight Vision-Language-Action model utilizing deep state space models to enable efficient, language-guided robotic manipulation on resource-constrained mobile platforms. By replacing standard transformer backbones with fast sequential state modeling, it achieves superior task success rates and inference speeds compared to larger counterparts like SmolVLA and $\pi_{0.5}$. This advancement addresses the critical deployment challenge of VLAs in real-world service robot environments without sacrificing performance.
AnolisVLA: Lightweight Vision-Language-Action Model with Deep State Space Models for Mobile Manipulation
arXiv:2603.15046v2 Announce Type: replace-cross Abstract: In this study, we address the problem of language-guided robotic manipulation, where a robot is required to manipulate a wide range of objects based on visual observations and natural language instructions. This task is essential for service robots that operate in human environments, and requires safety, efficiency, and task-level generality. Although Vision-Language-Action models (VLAs) have demonstrated strong performance for this task, their deployment in resource-constrained environments remains challenging because of the computational cost of standard transformer backbones. To overcome this limitation, we propose AnolisVLA, a lightweight VLA that uses a deep state space model to process multimodal sequences efficiently. The model uses lightweight, fast sequential state modeling to process visual and textual inputs. This design enables efficient trajectory generation. We evaluated the proposed method in simulation and on two real robot platforms. Compared with SmolVLA at a comparable parameter scale, AnolisVLA achieved higher task success rates, faster inference, and comparable GPU memory usage. In mobile manipulation, it outperformed the larger, 3-billion-parameter $\pi_{0.5}$ model by 21 percentage points in task success rate while achieving approximately three times faster inference.
Abstract:In this study, we address the problem of language-guided robotic manipulation, where a robot is required to manipulate a wide range of objects based on visual observations and natural language instructions. This task is essential for service robots that operate in human environments, and requires safety, efficiency, and task-level generality. Although Vision-Language-Action models (VLAs) have demonstrated strong performance for this task, their deployment in resource-constrained environments remains challenging because of the computational cost of standard transformer backbones. To overcome this limitation, we propose AnolisVLA, a lightweight VLA that uses a deep state space model to process multimodal sequences efficiently. The model uses lightweight, fast sequential state modeling to process visual and textual inputs. This design enables efficient trajectory generation. We evaluated the proposed method in simulation and on two real robot platforms. Compared with SmolVLA at a comparable parameter scale, AnolisVLA achieved higher task success rates, faster inference, and comparable GPU memory usage. In mobile manipulation, it outperformed the larger, 3-billion-parameter $\pi_{0.5}$ model by 21 percentage points in task success rate while achieving approximately three times faster inference.
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2603.15046 [cs.RO]
(or arXiv:2603.15046v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2603.15046 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 272 words
Related coverage
Just Noticeable Difference Modeling for Token Compression in Vision-Language-Action Models
arXiv · Aug 24, 2026
Build a Physical AI model factory with NVIDIA Cosmos 3 on SageMaker HyperPod
AWS · Sep 4, 2026
The Machines that Make the Machines
NVIDIA · Oct 7, 2026
NVIDIA Nemotron 3 Nano Omni Powers Multimodal Agent Reasoning in a Single Efficient Open Model
NVIDIA · Apr 28, 2026