# ActQuant: Sub-4-bit Action-Guided Quantization for Vision-Language-Action Models

arXiv: https://arxiv.org/abs/2605.24011 · HTML: https://mlsanity.com/paper/2605.24011/

## Key facts

- Cited by 8 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 35 of its 44 references that are in the map.
- First posted on arXiv 19 May 2026.

## Placed in

- Vision-language-action models / [Efficient VLA inference and adaptation](https://mlsanity.com/vla/efficient-vla-inference-and-adaptation/)

## Builds on (most cited first, 30 of 35)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers](https://mlsanity.com/paper/2210.17323/) (2022)
- [π0.5: a Vision-Language-Action Model with Open-World Generalization](https://mlsanity.com/paper/2504.16054/) (2025)
- [LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning](https://mlsanity.com/paper/2306.03310/) (2023)
- [AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration](https://mlsanity.com/paper/2306.00978/) (2023)
- [Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success](https://mlsanity.com/paper/2502.19645/) (2025)
- [New insights and perspectives on the natural gradient method](https://mlsanity.com/paper/1412.1193/) (2014)
- [CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulation](https://mlsanity.com/paper/2411.19650/) (2024)
- [SpQR: A Sparse-Quantized Representation for Near-Lossless LLM Weight Compression](https://mlsanity.com/paper/2306.03078/) (2023)
- [TinyVLA: Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulation](https://mlsanity.com/paper/2409.12514/) (2024)
- [SqueezeLLM: Dense-and-Sparse Quantization](https://mlsanity.com/paper/2306.07629/) (2023)
- [HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networks](https://mlsanity.com/paper/1911.03852/) (2019)
- [Extreme Compression of Large Language Models via Additive Quantization](https://mlsanity.com/paper/2401.06118/) (2024)
- [A Survey on Vision-Language-Action Models: An Action Tokenization Perspective](https://mlsanity.com/paper/2507.01925/) (2025)
- [EfficientVLA: Training-Free Acceleration and Compression for Vision-Language-Action Models](https://mlsanity.com/paper/2506.10100/) (2025)
- [QTIP: Quantization with Trellises and Incoherence Processing](https://mlsanity.com/paper/2406.11235/) (2024)
- [The HSIC Bottleneck: Deep Learning without Back-Propagation](https://mlsanity.com/paper/1908.01580/) (2019)
- [Self-Supervised Learning with Kernel Dependence Maximization](https://mlsanity.com/paper/2106.08320/) (2021)
- [SliM-LLM: Salience-Driven Mixed-Precision Quantization for Large Language Models](https://mlsanity.com/paper/2405.14917/) (2024)
- [Towards Mixed-Precision Quantization of Neural Networks via Constrained Optimization](https://mlsanity.com/paper/2110.06554/) (2021)
- [OMPQ: Orthogonal Mixed Precision Quantization](https://mlsanity.com/paper/2109.07865/) (2021)
- [Non-Structured DNN Weight Pruning -- Is It Beneficial in Any Platform?](https://mlsanity.com/paper/1907.02124/) (2019)
- [Quantization-Aware Imitation-Learning for Resource-Efficient Robotic Control](https://mlsanity.com/paper/2412.01034/) (2024)

## Built on by (most cited first, 8 of 8)

- [vla.cpp: A Unified Inference Runtime for Vision-Language-Action Models](https://mlsanity.com/paper/2606.08094/) (2026)
- [Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots](https://mlsanity.com/paper/2607.02501/) (2026)
- [PhyAI: Real-Time Physical AI at the Edge, Scalable Rollouts in the Cloud](https://mlsanity.com/paper/2608.03682/) (2026)
- [Bit-Flip Attacks on Vision-Language-Action Models: Action-Decoding Architecture Shapes the Vulnerability](https://mlsanity.com/paper/2608.15475/) (2026)
- [FoldQuantVLA: Native Low-Bit Quantization of Vision-Language-Action Models via Consistent Folding](https://mlsanity.com/paper/2609.24433/) (2026)
- [Q-WAM: 4-Bit Quantization of World Action Models with Action-Subspace Protection](https://mlsanity.com/paper/2609.33269/) (2026)
- [SteerQuant: Steering Quantization Error with Action-Guided Scaling in World-Action Models](https://mlsanity.com/paper/2609.39056/) (2026)
- [ActTune: Action-Aware Precision and GPU Operating-Point Adaptation for Energy-Efficient Vision-Language-Action Inference](https://mlsanity.com/paper/2610.08444/) (2026)

## Abstract (arXiv)

Vision-Language-Action (VLA) models exhibit remarkable action generation for embodied intelligence, but their heavy compute make deployment on edge platforms impractical. Aggressive, sub-4-bit weight quantization is the natural solution, yet existing post-training quantization (PTQ) methods suffer severe performance degradation in this regime. To address this, we introduce ActQuant, an action-guided mixed-precision PTQ framework that operates in two stages: (1) an inter-tensor bit allocator that assigns each weight matrix a single bit-width based on how much it contributes to predicting the agent's actions; (2) an intra-tensor scale optimizer tunes per-block quantization scales using action-aware curvature, so that dynamic range is concentrated on the weights most influential for control. To deliver the on-device benefits of our aggressive quantization, we further introduce OmniModel.cpp, an agentic conversion pipeline that ports architectures into a native C/C++ runtime with efficient low-bit kernels. We evaluate ActQuant both in simulation and on a real-world 6-DoF UR3 arm, with all models deployed through OmniModel.cpp. On the LIBERO benchmark, ActQuant is the only method that operates at or below 3 bits-per-weight, retaining 95.0% on OpenVLA-OFT and 94.8% on π0.5. Pushed further, ActQuant reaches 2.5 bpw at 90.1% on OpenVLA-OFT, compressing the backbone from 14.3 GB to 2.7 GB (5.3×). On the physical UR3 arm, π0.5 quantized with ActQuant retains the baseline's success rate while reducing the memory footprint by 2.5×.