mlsanityData to 13 Oct 2026

Running VLAs at Real-time Speed

arXiv 2510.26742 ↗First posted 30 Oct 2025cs.ROcited by 43

Not on the map. Why?

Cited by 36 in Vision-language-action models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

In this paper, we show how to run pi0-level multi-view VLA at 30Hz frame rate and at most 480Hz trajectory frequency using a single consumer GPU. This enables dynamic and real-time tasks that were previously believed to be unattainable by large VLA models. To achieve it, we introduce a bag of strategies to eliminate the overheads in model inference. The real-world experiment shows that the pi0 policy with our strategy achieves a 100% success rate in grasping a falling pen task. Based on the results, we further propose a full streaming inference framework for real-time robot control of VLA. Code is available at https://github.com/Dexmal/realtime-vla.

Key facts. Cited by 43 papers in the map (data through 13 Oct 2026). First posted on arXiv 30 Oct 2025.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2025: 3; 2026: 40.

43 papers.

Sort:

By area:

  1. 2025VLASH: Real-Time VLAs via Future-State-Aware Asynchronous Inferencecited by 56
  2. 2026Xiaomi-Robotics-0: An Open-Sourced Vision-Language-Action Model with Real-Time Executioncited by 44
  3. 2026DynamicVLA: A Vision-Language-Action Model for Dynamic Object Manipulationcited by 33
  4. 2025VLA-AN: An Efficient and Onboard Vision-Language-Action Framework for Aerial Navigation in Complex Environmentscited by 23
  5. 2026FASTER: Rethinking Real-Time Flow VLAscited by 21
  6. 2025VLA-RAIL: A Real-Time Asynchronous Inference Linker for VLA Models and Robotscited by 18
  7. 2026How Fast Can I Run My VLA? Demystifying VLA Inference Performance with VLA-Perfcited by 18
  8. 2026FutureRTC: Real-Time Robot Execution with Anticipatory-Conditioned Action Chunkingcited by 10
  9. 2026Shallow-π: Knowledge Distillation for Flow-based VLAscited by 8
  10. 2026GuidedVLA: Specifying Task-Relevant Factors via Plug-and-Play Action Attention Specializationcited by 8
  11. 2026Realtime-VLA FLASH: Speculative Inference Framework for Diffusion-based VLAscited by 8
  12. 2026Characterizing Vision-Language-Action Models across XPUs: Constraints and Acceleration for On-Robot Deploymentcited by 6
31 more
  1. 2026TIDAL: Temporally Interleaved Diffusion and Action Loop for High-Frequency VLA Controlcited by 5
  2. 2026FLARE: Diffusion for Hybrid Language Modelcited by 5
  3. 2026OxyGen: Unified KV Cache Management for VLA Inference under Multi-Task Parallelismcited by 4
  4. 2026FlashVLA: Streaming Action Decoding for Fast and Asynchronous VLA Inferencecited by 4
  5. 2026Speedup Patch: Learning a Plug-and-Play Policy to Accelerate Embodied Manipulationcited by 3
  6. 2026DiscreteRTC: Discrete Diffusion Policies are Natural Asynchronous Executorscited by 3
  7. 2026Understanding Asynchronous Inference Methods for Vision-Language-Action Modelscited by 3
  8. 2026TempoVLA: Learning Speed-Controllable Vision-Language-Action Policiescited by 3
  9. 2026Jetson-PI: Towards Onboard Real-Time Robot Control via Foresight-Aligned Asynchronous Inferencecited by 3
  10. 2026Reflex: Real-Time VLA Control through Streaming Inferencecited by 3
  11. 2026From Inference Efficiency to Embodied Efficiency: Revisiting Efficiency Metrics for Vision-Language-Action Modelscited by 2
  12. 2026VLK: Learning Humanoid Loco-Manipulation from Synthetic Interactions in Reconstructed Scenescited by 2
  13. 2026DynamicManip: Enabling Dynamic Manipulation from a Single Static Demonstrationcited by 2
  14. 2026Reinforcement Learning for Real-Time Vision-Language-Action Policiescited by 2
  15. 2026Environment-Aware Adaptive Pruning with Interleaved Inference Orchestration for Vision-Language-Action Modelscited by 1
  16. 2026RIO: Flexible Real-Time Robot I/O for Cross-Embodiment Robot Learningcited by 1
  17. 2026PhyAI: Real-Time Physical AI at the Edge, Scalable Rollouts in the Cloudcited by 1
  18. 2026Reflex: Enabling Fast and Predictive Vision-Language-Action Models for Reaction-Critical Manipulationcited by 1
  19. 2026EcoVLA: Energy-Efficient Device-Edge Co-Inference for Vision-Language-Action Models under Real-Time Constraintscited by 1
  20. 2026FluxVLA Engine: A One-Stop VLA Engineering Platform for Embodied Intelligencecited by 1
  21. 20262D or 3D: Who Governs Salience in VLA Models? -- Tri-Stage Token Pruning Framework with Modality Salience Awarenesscited by 0
  22. 2026LiteVLA-H: Dual-Rate Vision-Language-Action Inference for Onboard Aerial Guidance and Semantic Perceptioncited by 0
  23. 2026Assistron: Bayesian Shared Autonomy with Off-the-shelf Vision-Language-Action Modelscited by 0
  24. 2026PonderPounce: A Pretrained MLLM as an Episode Context Engine for Robot Controlcited by 0
  25. 2026IMLE-VLA: Fast Single-Step Action Generation for Vision-Language-Action Policiescited by 0
  26. 2026Efficient Vision-Language-Action Management and Serving for Robot Factoriescited by 0
  27. 2026rMuscle: Robotic Muscle Memory for Efficient Vision-Language-Action Model Inferencecited by 0
  28. 2026KerColle: Unlocking Fine-Grained GPU Concurrency in Vision-Language-Action Modelscited by 0
  29. 2026RAVEL: Asynchronous Rolling Inference for Flow-Based Vision-Language-Action Modelscited by 0
  30. 2026Toward Real-Time VLAs: Stage-Aware Two-Step Flow Denoising and System-Level Evaluationcited by 0
  31. 2026REACT: Rolling Denoising and Dual Decoupling for Reactive Robot Control with VLA Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Realtime-VLA V2: Learning to Run VLAs Fast, Smooth, and Accurate
  2. 2026vla.cpp: A Unified Inference Runtime for Vision-Language-Action Models
  3. 2026PiL-World: A Chunk-Wise World Model for VLA Policy-in-the-Loop Evaluation
  4. 2025SP-VLA: A Joint Model Scheduling and Token Pruning Approach for VLA Model Acceleration
  5. 2026Deltoris: Enabling Real-time VLA Inference in Embodied AI via Bit-level Sparsity and Speculative Inference
  6. 2026A Motion-Aware Vector Quantization Framework with Centroid Reuse for Efficient VLA Inference
  7. 2025Cross-Platform Scaling of Vision-Language-Action Models from Edge to Cloud GPUs
  8. 2026SimVLA: A Simple VLA Baseline for Robotic Manipulation