aiqbench analysis
This research addresses a critical bottleneck in Vision Language Action Models (VLAs) by introducing MiGA, a dataset that explicitly trains models on diverse gripper types rather than assuming uniform interaction strategies. The proposed GVLA framework utilizes adapter-based policy routing and multi-gripper tokenization to balance parameter efficiency with the distinct physical requirements of different end-effectors. These advances significantly improve zero-shot generalization for real-world robotic manipulation tasks across varying embodiments.
Gripper-aware Vision Language Action Models
arXiv:2608.24603v1 Announce Type: new Abstract: Vision language action models (VLAs) have advanced general purpose robotic grasping and manipulation by enabling robots to interpret visual observations and natural language instructions to generate executable action sequences. However, existing VLAs often implicitly assume gripper invariance, despite grasping strategies being inherently embodiment-dependent. Different gripper types, such as parallel-jaw and suction, usually require distinct interaction strategies to achieve the same grasping objective. Moreover, current datasets for VLAs predominantly rely on parallel-jaw grippers, limiting gripper-aware learning. To address this gap, we introduce MiGA, a multi-gripper-aware dataset spanning five distinct gripper types across multiple robots with 103,000 demonstrations, explicitly capturing strategy divergence under shared task objectives. We further propose GVLA, which combines a new multi-gripper tokenizer with adapter-based policy routing. Our new gripper encoding induces structured embedding information that balances parameter sharing and strategy differentiation, while layer-wise probing confirms meaningful gripper-conditioned representations for VLAs. Intensive experiments in both simulation and real-world robots show that our GVLA outperforms the current baselines across evaluated settings. Our method also improves zero-shot generalization or few-shot adaptation to new objects or unseen tasks, and enable more efficient gripper adaptation.
Authors:Hanyi Zhang, Zihong Luo, Tianyu Li, Khang Nguyen, Basu Hela, Shreyas Kumar, Ngoc Duy Tran, Feng Dai, Charith Munasinghe, Jorge Peña Queralta, Giovanni Toffetti, Khoa Vo, Ngan Le, Ravi Prakash, Quan Vuong, Tung D. Ta, Long Hu, Anh Nguyen, Baoru Huang
Abstract:Vision language action models (VLAs) have advanced general purpose robotic grasping and manipulation by enabling robots to interpret visual observations and natural language instructions to generate executable action sequences. However, existing VLAs often implicitly assume gripper invariance, despite grasping strategies being inherently embodiment-dependent. Different gripper types, such as parallel-jaw and suction, usually require distinct interaction strategies to achieve the same grasping objective. Moreover, current datasets for VLAs predominantly rely on parallel-jaw grippers, limiting gripper-aware learning. To address this gap, we introduce MiGA, a multi-gripper-aware dataset spanning five distinct gripper types across multiple robots with 103,000 demonstrations, explicitly capturing strategy divergence under shared task objectives. We further propose GVLA, which combines a new multi-gripper tokenizer with adapter-based policy routing. Our new gripper encoding induces structured embedding information that balances parameter sharing and strategy differentiation, while layer-wise probing confirms meaningful gripper-conditioned representations for VLAs. Intensive experiments in both simulation and real-world robots show that our GVLA outperforms the current baselines across evaluated settings. Our method also improves zero-shot generalization or few-shot adaptation to new objects or unseen tasks, and enable more efficient gripper adaptation.
Robotics (cs.RO)
arXiv:2608.24603 [cs.RO]
(or arXiv:2608.24603v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2608.24603 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 240 words
Related coverage
YUBI-STAG: Contact and Semantic-Rich Alignment for VLAs via Automated Video-Language Grounding
arXiv · Oct 8, 2026
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · Oct 13, 2026
AnchorRefine: Synergy-Manipulation Based on Trajectory Anchor and Residual Refinement for Vision-Language-Action Models
arXiv · Jul 23, 2026
World-Calibrated Proposal-to-Action Flow for Vision-Language-Action Models
arXiv · Oct 5, 2026