mlsanityData to 14 Oct 2026

EfficientVLA: Training-Free Acceleration and Compression for Vision-Language-Action Models

arXiv 2506.10100 ↗First posted 11 Jun 2025cs.CVcited by 93

On the map: Vision-language-action models → Efficient VLA inference and adaptation

also in 1 more
Abstract (arXiv)

Vision-Language-Action (VLA) models, particularly diffusion-based architectures, demonstrate transformative potential for embodied intelligence but are severely hampered by high computational and memory demands stemming from extensive inherent and inference-time redundancies. While existing acceleration efforts often target isolated inefficiencies, such piecemeal solutions typically fail to holistically address the varied computational and memory bottlenecks across the entire VLA pipeline, thereby limiting practical deployability. We introduce EfficientVLA, a structured and training-free inference acceleration framework that systematically eliminates these barriers by cohesively exploiting multifaceted redundancies. EfficientVLA synergistically integrates three targeted strategies: (1) pruning of functionally inconsequential layers from the language module, guided by an analysis of inter-layer redundancies; (2) optimizing the visual processing pathway through a task-aware strategy that selects a compact, diverse set of visual tokens, balancing task-criticality with informational coverage; and (3) alleviating temporal computational redundancy within the iterative diffusion-based action head by strategically caching and reusing key intermediate features. We apply our method to a standard VLA model CogACT, yielding a 1.93X inference speedup and reduces FLOPs to 28.9%, with only a 0.6% success rate drop in the SIMPLER benchmark.

Key facts. Cited by 93 papers in the map (data through 14 Oct 2026). Placed in Efficient VLA inference and adaptation, one of 13 directions in Vision-language-action models. Builds on 43 of its 52 references that are in the map. First posted on arXiv 11 Jun 2025.

Builds on

43 of 52 references.

Closest ancestors · 2

Same direction, newest first

  1. 2025MoLe-VLA: Dynamic Layer-skipping Vision Language Action Model via Mixture-of-Layers for Efficient Robot Manipulation26 Mar
  2. 2025PD-VLA: Accelerating Vision-Language-Action Model Integrated with Action Chunking via Parallel Decoding4 Mar

Other references · 41

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2023GPT-4 Technical Reportcited by 18,396
  3. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  4. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  5. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  6. 2023Visual Instruction Tuningcited by 7,869
  7. 2022Flamingo: a Visual Language Model for Few-Shot Learningcited by 4,386
  8. 2022BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generationcited by 4,322
  9. 2023Sigmoid Loss for Language Image Pre-Trainingcited by 3,109
  10. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  11. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 2,887
  12. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 2,407
29 more
  1. 2023LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learningcited by 1,410
  2. 2023Open X-Embodiment: Robotic Learning Datasets and RT-X Modelscited by 862
  3. 2024An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Modelscited by 642
  4. 2023OpenFlamingo: An Open-Source Framework for Training Large Autoregressive Vision-Language Modelscited by 565
  5. 2024Evaluating Real-World Robot Manipulation Policies in Simulationcited by 495
  6. 2024CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulationcited by 422
  7. 2024LLaVA-PruMerge: Adaptive Token Reduction for Efficient Large Multimodal Modelscited by 402
  8. 2024SparseVLM: Visual Token Sparsification for Efficient Vision-Language Model Inferencecited by 391
  9. 2023Vision-Language Foundation Models as Effective Robot Imitatorscited by 349
  10. 2024Prismatic VLMs: Investigating the Design Space of Visually-Conditioned Language Modelscited by 339
  11. 2024TinyVLA: Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulationcited by 339
  12. 2025DexVLA: Vision-Language Model with Plug-In Diffusion Expert for General Robot Controlcited by 195
  13. 2023RH20T: A Comprehensive Robotic Dataset for Learning Diverse Skills in One-Shotcited by 183
  14. 2024Accelerating Diffusion Transformers with Token-wise Feature Cachingcited by 156
  15. 2025Stop Looking for Important Tokens in Multimodal Language Models: Duplication Matters Morecited by 155
  16. 2024FORA: Fast-Forward Caching in Diffusion Transformer Accelerationcited by 137
  17. 2024DeeR-VLA: Dynamic Inference of Multimodal Large Language Models for Efficient Robot Executioncited by 117
  18. 2024RoboMamba: Efficient Vision-Language-Action Model for Robotic Reasoning and Manipulationcited by 103
  19. 2025OpenHelix: A Short Survey, Empirical Analysis, and Open-Source Dual-System VLA Model for Robotic Manipulationcited by 82
  20. 2024Multi-Stage Vision Token Dropping: Towards Efficient Multimodal Large Language Modelcited by 68
  21. 2022EfficientVLM: Fast and Accurate Vision-Language Models via Knowledge Distillation and Modal-adaptive Pruningcited by 56
  22. 2024MC-LLaVA: Multi-Concept Personalized Vision-Language Modelcited by 54
  23. 2024LLM as Dataset Analyst: Subpopulation Structure Discovery with Large Language Modelcited by 44
  24. 2024SlimGPT: Layer-wise Structured Pruning for Large Language Modelscited by 44
  25. 2025Shifting AI Efficiency From Model-Centric to Data-Centric Compressioncited by 29
  26. 2025Data Whisperer: Efficient Data Selection for Task-Specific LLM Fine-Tuning via Few-Shot In-Context Learningcited by 20
  27. 2025You Only Prune Once: Designing Calibration-Free Model Compression With Policy Learningcited by 17
  28. 2024Why Lift so Heavy? Slimming Large Language Models by Cutting Off the Layerscited by 2
  29. 2024TrimLLM: Progressive Layer Dropping for Domain-Specific LLMscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 25; 2026: 68.

93 papers.

Sort:

By area:

  1. 2025Large VLM-based Vision-Language-Action Models for Robotic Manipulation: A Surveycited by 69
  2. 2025SP-VLA: A Joint Model Scheduling and Token Pruning Approach for VLA Model Accelerationcited by 58
  3. 2026InfiniteVGGT: Visual Geometry Grounded Transformer for Endless Streamscited by 56
  4. 2025HAMLET: Switch your Vision-Language-Action Model into a History-Aware Policycited by 46
  5. 2025SpecPrune-VLA: Accelerating Vision-Language-Action Models via Action-Aware Self-Speculative Pruningcited by 43
  6. 2025The Better You Learn, The Smarter You Prune: Towards Efficient Vision-language-action Models via Differentiable Token Pruningcited by 36
  7. 2025Action-aware Dynamic Pruning for Efficient Vision-Language-Action Manipulationcited by 34
  8. 2025A Survey on Efficient Vision-Language-Action Modelscited by 34
  9. 2026QuantVLA: Scale-Calibrated Post-Training Quantization for Vision-Language-Action Modelscited by 30
  10. 2026QVLA: Not All Channels Are Equal in Vision-Language-Action Model's Quantizationcited by 28
  11. 2025Efficient Vision-Language-Action Models for Embodied Manipulation: A Systematic Surveycited by 26
  12. 2025RLRC: Reinforcement Learning-based Recovery for Compressed Vision-Language-Action Modelscited by 23
38 more
  1. 2025Efficient Multi-modal Large Language Models via Progressive Consistency Distillationcited by 22
  2. 2025AudioMarathon: A Comprehensive Benchmark for Long-Context Audio Understanding and Efficiency in Audio LLMscited by 22
  3. 2026FASTER: Rethinking Real-Time Flow VLAscited by 21
  4. 2025Are We Using the Right Benchmark: An Evaluation Framework for Visual Token Compression Methodscited by 17
  5. 2025Bridging the Semantic-Action Gap in Visual Token Pruning for Efficient VLA Inferencecited by 17
  6. 2025Prune2Drive: A Plug-and-Play Framework for Accelerating Vision-Language Models in Autonomous Drivingcited by 16
  7. 2025SQAP-VLA: A Synergistic Quantization-Aware Pruning Framework for High-Performance Vision-Language-Action Modelscited by 15
  8. 2026SnapFlow: One-Step Action Generation for Flow-Matching VLAs via Progressive Self-Distillationcited by 14
  9. 2025NanoVLA: Routing Decoupled Vision-Language Understanding for Nano-sized Generalist Robotic Policiescited by 12
  10. 2026DyQ-VLA: Temporal-Dynamic-Aware Quantization for Embodied Vision-Language-Action Modelscited by 12
  11. 2025Which Heads Matter for Reasoning? RL-Guided KV Cache Compressioncited by 10
  12. 2026ABPolicy: Asynchronous B-Spline Flow Policy for Real-Time and Smooth Robotic Manipulationcited by 10
  13. 2026TurboVLA: Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAMcited by 10
  14. 2026Innovator-VL: A Multimodal Large Language Model for Scientific Discoverycited by 8
  15. 2026Shallow-π: Knowledge Distillation for Flow-based VLAscited by 8
  16. 2026A1: A Fully Transparent Open-Source, Adaptive and Efficient Truncated Vision-Language-Action Modelcited by 8
  17. 2026Realtime-VLA FLASH: Speculative Inference Framework for Diffusion-based VLAscited by 8
  18. 2026ActQuant: Sub-4-bit Action-Guided Quantization for Vision-Language-Action Modelscited by 8
  19. 2025Survey of Vision-Language-Action Models for Embodied Manipulationcited by 7
  20. 2025Compressor-VLA: Instruction-Guided Visual Token Compression for Efficient Robotic Manipulationcited by 7
  21. 2025Token Expand-Merge: Training-Free Token Compression for Vision-Language-Action Modelscited by 7
  22. 2025TS-DP: Reinforcement Speculative Decoding For Temporal Adaptive Diffusion Policy Accelerationcited by 6
  23. 2026AC^2-VLA: Action-Context-Aware Adaptive Computation in Vision-Language-Action Models for Efficient Robotic Manipulationcited by 6
  24. 2026StreamingVLA: Streaming Vision-Language-Action Model with Action Flow Matching and Adaptive Early Observationcited by 6
  25. 2025ActDistill: General Action-Guided Self-Derived Distillation for Efficient Vision-Language-Action Modelscited by 5
  26. 2025IPCV: Information-Preserving Compression for MLLM Visual Encoderscited by 5
  27. 2026Sparse ActionGen: Accelerating Diffusion Policy with Real-time Pruningcited by 5
  28. 2026TIDAL: Temporally Interleaved Diffusion and Action Loop for High-Frequency VLA Controlcited by 5
  29. 2026Efficient Long-Horizon Vision-Language-Action Models via Static-Dynamic Disentanglementcited by 5
  30. 2026EvoStreaming: Your Offline Video Model Is a Natively Streaming Assistantcited by 5
  31. 2026Recursive Belief Vision Language Action Modelscited by 4
  32. 2026OxyGen: Unified KV Cache Management for VLA Inference under Multi-Task Parallelismcited by 4
  33. 2026See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Modelcited by 4
  34. 2026BFA++: Hierarchical Best-Feature-Aware Token Prune for Multi-View Vision Language Action Modelcited by 3
  35. 2026ProbeFlow: Training-Free Adaptive Flow Matching for Vision-Language-Action Modelscited by 3
  36. 2026Speedup Patch: Learning a Plug-and-Play Policy to Accelerate Embodied Manipulationcited by 3
  37. 2026SAFE-Pruner: Semantic Attention-Guided Future-Aware Token Pruning for Efficient Vision-Language-Action Manipulationcited by 3
  38. 2026TempoVLA: Learning Speed-Controllable Vision-Language-Action Policiescited by 3

50 most cited of 93 All 93 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Think Twice, Act Once: Token-Aware Compression and Action Reuse for Efficient Inference in Vision-Language-Action Models
  2. 2026Recurrent-Depth VLA: Implicit Test-Time Compute Scaling of Vision-Language-Action Models via Latent Iterative Reasoning
  3. 2025KV-Efficient VLA: A Method to Speed up Vision Language Models with RNN-Gated Chunked KV Cache
  4. 2026DAM-VLA: Decoupled Asynchronous Multimodal Vision Language Action model
  5. 2025Discrete Diffusion VLA: Bringing Discrete Diffusion to Action Decoding in Vision-Language-Action Policies
  6. 2026FocusVLA: Focused Visual Utilization for Vision-Language-Action Models
  7. 2026What Makes an Efficient VLA? Navigating Action-Head Design, Scaling, and Latency
  8. 2026UAM: A Dual-Stream Perspective on Forgetting in VLA Training