Efficient VLA Models

VLA: Vision-Language-Action

Efficient VLA research studies how robots and other embodied agents can connect visual observations and language instructions to actions under limited compute, memory, and latency budgets. Common approaches include compact backbones and action heads, visual-token pruning, quantization and cache reuse, and faster action generation through chunking, parallel decoding, or fewer diffusion steps. Other methods adapt computation to the task or reuse compressed memory so models can respond quickly while retaining context for long-horizon manipulation, navigation, and driving. Results on benchmarks such as LIBERO, RoboCasa, and driving and real-robot evaluations show that many methods can improve speed or reduce resource use, but efficiency must also be judged by task success, robustness, motion quality, and end-to-end execution cost.

Latest papers

106 papers17 papers in the last four weeks, up 13% on the four weeks before. 0.1% of all new papers.

OpenFlyV2: An onboard vision-language-action model for aerial perception and navigation
Oct 29, 2026Junli Liu, Zhongrui You, Chenhui Li +9Aerial VLN Vision-Language-Action Models
Shanghai Artificial Intelligence Laboratory · Northwestern Polytechnical University · Beihang University +4
Quantize by Doing: Task-Guided Mixed-Precision Quantization of Vision-Language-Action Models
Oct 28, 2026Jiuyi Xu, Xiao Hu, Meida Chen +3Mixed-Precision Quantization VLM Quantization
Colorado School of Mines · Northeastern University · USC Institute for Creative Technologies +1
Vision-Language-Action or World-Action? Just Action Models
Oct 27, 2026Yonghao Yu, Lang Huang, Runyi Li +1VLA Model Training World Action Models
The University of Tokyo · National Institute of Informatics
Full-Weight VLA Fine-Tuning Without an fp32 Master Copy Silently Erases Most Updates
Oct 22, 2026Shubham SinghVLA Model Adaptation Efficient VLA Models
SmolVLA-MoE: Sparse Mixture-of-Experts for Efficient Multi-Task Robotic Manipulation
Oct 22, 2026Shiyang Zhang, Miao Yang, Zhengda Li +4Efficient VLA Models Vision-Language-Action Models
College of Informatics, Huazhong Agricultural University, No. 1 Shizi Mountain Street, Hongshan District, Wuhan 430070, China · College of Plant Science and Technology, Huazhong Agricultural University, No. 1 Shizi Mountain Street, Hongshan District, Wuhan 430070, China · Hubei Hongshan Laboratory, Wuhan 430070, China
CondenseVLA: Learnable History Condensation for Efficient Multi-Frame VLA
Oct 19, 2026Feiyang Hong, Yujie Wei, Bo Zhao +3Vision-Language-Action Models Efficient VLA Models
Harbin Institute of Technology, Shenzhen · Harbin Institute of Technology · Shanghai Jiao Tong University +1
Think Densely, Act Sparsely: Latent Expert Cognitive Chains for Vision-Language-Action Autonomous Driving
Oct 19, 2026Jie Wang, Guang Li, Zhijian Huang +5VLMs for Autonomous Driving Efficient VLA Models
School of Artificial Intelligence, Tianjin University · School of Artificial Intelligence, Shanxi University · Xiaomi EV
NestedVLA: Learning to Consolidate and Generate Skills for Vision-Language-Action Models
Oct 16, 2026Guodong Du, Wanyu LinVision-Language-Action Models Continual Learning for VLA Models
Department of Computing, The Hong Kong Polytechnic University
REKI: Registering Knowledge in VLAs
Oct 15, 2026Juliette Marrie, Paul Pacaud, Adrien Ramanana Rahary +2Vision-Language-Action Models Efficient VLA Models
Kyutai · Inria · École nationale des ponts et chaussées
Vision-Predictor-Action: Towards Efficient Embodied AI via Neuro-Symbolic JEPA Predictors
Oct 11, 2026Yeonseok LeeFlow Matching for Action Generation Efficient VLA Models
SLING AI Inc., Incheon, Republic of Korea
Evo-World: A Lightweight Vision-Language-Action Model with Future Motion Prediction via Optical-Flow Supervision
Oct 9, 2026Zile Zhong, Yuxin Du, Tao Lin +4Vision-Language-Action Models Efficient VLA Models
School of AI, Shanghai Jiao Tong University · Zhejiang University · University of Electronic Science and Technology of China +1
Tell Robot What Not to Do: A Negation Understanding Perspective
Oct 8, 2026Fazeng Li, Gan Sun, Hao Cheng +2Vision-Language-Action Models Reasoning for VLA Models
South China University of Technology · Harbin Institute of Technology
NavGPT-3: Harnessing Context in a Hierarchical Navigation Runtime
Oct 7, 2026Gengze Zhou, Yicong Hong, Jiazhao Zhang +9Robot Navigation Vision-Language Navigation
AIML, Adelaide University · Metacognition · Roblox +5
VLA-ACL: Action-Consistent Visual Token Pruning for Efficient Vision-Language-Action Models
Oct 6, 2026Owen Du, Yang Yue, Jie Zhang +3Efficient VLA Models VLMs for Robotics
ETH Zurich · Tsinghua University
GeoScaffold: Learning Compact Geometric Latents via Reconstruction for Efficient Vision-Language Navigation
Oct 2, 2026Yixuan Jiang, Wentong Li, An Liu +5Vision-Language Navigation Efficient VLA Models
Nanjing University, Nanjing, China · Institute of Automation, Chinese Academy of Sciences, Beijing, China · Nanjing University of Aeronautics and Astronautics, Nanjing, China +1
eRLT: Efficient VLA Reinforcement Learning via Action-Relevant Token Routing
Oct 1, 2026Dehao Huang, Jianbang Liu, Jianpan Gao +7Efficient VLA Models Robotic Manipulation
Southern University of Science and Technology, Shenzhen, China. · Beijing Zhongguancun Academy, Beijing, China. · Samsung Robotics eXperience. +2
Spike-driven Vision-Language-Action Model
Sep 30, 2026Shuai Wang, Malu Zhang, Mingquan Liu +6Efficient VLA Models Vision-Language-Action Models
University of Electronic Science and Technology of China.
ActionUNet: Improving Robustness of VLA Models with Efficient Multi-scale Fine-tuning
Sep 28, 2026Di Zhu, Ziheng Yan, Fang WanRobotic Control Efficient VLA Models
University of Chinese Academy of Sciences
Adjoint Guidance Flow: Amortized Critic Guidance for VLA Policies
Sep 28, 2026Jeongsol Kim, Youngjun Jun, Kyumin Choi +6Efficient VLA Models Vision-Language-Action Models
KAIST · SKKU · GIST
FailPatch: Failure Residual Patching for Vision-Language-Action Models
Sep 28, 2026Peng Yu, Jiacheng Wang, Ziheng Zhang +6Robot Policy Learning Robot Policy Adaptation
Xi’an Jiaotong University · Dexmal · Nanjing University
SLIP-VLA: Single-Step Latent Imagination for Policy Learning in Vision-Language-Action Models
Sep 27, 2026Tianfu Li, Haoxuan Xu, Wenbo Chen +7World Model Learning Visuomotor Policy Learning
The Hong Kong University of Science and Technology (Guangzhou). · The Hong Kong University of Science and Technology. · Nanyang Technological University. +1
Towards VLA-Dreamer: Refining VLA Behavior Using World Models
Sep 25, 2026Parsa Mastouri Kashani, Jan-Gerrit Habekost, Stefan WermterWorld Models for VLA Models Vision-Language-Action Models
Knowledge Technology, Dept. of Informatics, University of Hamburg, Germany
FRAM: Trajectory-Guided Visual Feature Selection for Compact Language-Conditioned Robot Manipulation
Sep 25, 2026Hiroshi Ito, Hyogo Hiruma, Yoshiki Kanai +3Language-Conditioned Robot Manipulation Efficient VLA Models
Research & Development Group, Hitachi, Ltd., Ibaraki, 312-0034, Japan.
VLALight: Lightweight Vision-Language-Action Models for Emergency-Aware Traffic Signal Control
Sep 25, 2026Kemou Jiang, Maonan Wang, Xingchen Zou +6Traffic Signal Control Efficient VLA Models
School of Transportation Science and Engineering, Beihang University, Beijing, China · The Chinese University of Hong Kong (Shenzhen), Shenzhen, China · The Chinese University of Hong Kong, Hong Kong SAR, China +2
TANDEM: Task and Motion Planning with As-Needed Demonstrations for Efficient Vision-Language-Action Model Fine-tuning
Sep 23, 2026Samrat Sahoo, Liang Ji, Tom Silver +1Data-Efficient Robot Learning Long-Horizon Robotic Manipulation
Stanford University. · Princeton University.
SmoLSTM: A Compact Vision-Language-Action Model with Recurrent Memory that Persists
Sep 19, 2026Jan-Gerrit Habekost, Parsa Mastouri Kashani, Connor Gäde +5Flow Matching for Action Generation Vision-Language-Action Models
Knowledge Technology Group, Department of Informatics, University of Hamburg, Hamburg, Germany
Recovering Aggressively Pruned Vision-Language-Action Models with Offline Hidden-State Distillation
Sep 17, 2026Chiyoung Kim, Sanghyuk Roy Choi, Minhyeok LeeEfficient VLA Models Vision-Language-Action Models
Chung-Ang University, Seoul, Republic of Korea
FIVE-VLA: Fast and EffectIVE Autonomous Driving with Recurrent Action Memory
Sep 16, 2026Kemal Oksuz, Alexandru Buburuzan, Yuhan Yao +1Memory-Augmented VLMs VLMs for Autonomous Driving
Robert Bosch GmbH, Germany · Five AI Ltd., United Kingdom
Decoupling Vision, Language, and Action for Efficient Multi-Task Robot Policies
Sep 16, 2026Xiatao Sun, Chen Liang, Ziyao Zeng +5Visuomotor Policy Learning Efficient VLA Models
Department of Computer Science, Yale University, New Haven, CT, USA. · Peking University, Beijing, China. · Digients, Singapore.
Dense to MoE Adaptation for Compact Vision Language Action Policies
Sep 15, 2026Muchun Niu, Shuang Chen, Yuzhou Wu +5Robot Policy Adaptation Efficient VLA Models
Shanghai Jiao Tong University, Shanghai, China · Tianji KernalMind Co., Ltd.

Oct 29, 2026J/K move · Enter open · S save

OpenFlyV2: An onboard vision-language-action model for aerial perception and navigation

Junli Liu, Zhongrui You, Chenhui Li, Zhen Li+8

Shanghai Artificial Intelligence Laboratory · Northwestern Polytechnical University · Beihang University+4

Aerial VLN Vision-Language-Action Models Efficient VLA Models Closed-Loop Control of VLA Models Spatial Grounding for VLA Models

Open paper AskPDF arXiv

Human cognition flexibly draws on a shared perceptual foundation to understand environments, attend to referenced entities, reason about spatial relationships, and act. Endowing aerial robots with an analogous functional organization remains challenging, as existing systems are often specialized for individual capabilities, dependent on offboard computation, or unable to sustain closed-loop operation on resource-constrained platforms. We present OpenFlyV2, a unified onboard vision-language-action architecture that uses a shared vision-language representation to support semantic understanding, language-grounded perception, spatial reasoning, and continuous aerial control within a single model. A compact action expert converts the shared multimodal context into short-horizon flight waypoints, while motion-aware temporal memory and action priors reduce the computational cost of closed-loop execution. We construct \dataset{}, a simulation-based dataset comprising approximately 70,000 navigation trajectories and 60,000 vision-language instruction--response pairs. OpenFlyV2 generates action chunks, each comprising 12 waypoints, entirely onboard at 10 Hz using an NVIDIA Jetson Orin NX. It transfers zero-shot from simulation to physical quadrotor flight, achieving a 48% navigation success rate across four real-world environments without real-world aerial training data. Extensive results demonstrate that broad vision-language competence and continuous closed-loop flight can coexist within a single model on a resource-constrained aerial platform. Code will be made publicly available at https://github.com/Ideal-ljl/AeroBrain.