# 2D or 3D: Who Governs Salience in VLA Models? -- Tri-Stage Token Pruning Framework with Modality Salience Awareness

arXiv: https://arxiv.org/abs/2604.09244 · HTML: https://mlsanity.com/paper/2604.09244/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in VLA architectures and generalization, one of 13 directions in Vision-language-action models.
- Builds on 30 of its 46 references that are in the map.
- First posted on arXiv 10 Apr 2026.

## Placed in

- Vision-language-action models / [VLA architectures and generalization](https://mlsanity.com/vla/vla-architectures-and-generalization/)
- Efficient LLM inference and serving / [Visual token pruning in multimodal models](https://mlsanity.com/inference/visual-token-pruning-in-multimodal-models/)

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

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [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)
- [PaLM-E: An Embodied Multimodal Language Model](https://mlsanity.com/paper/2303.03378/) (2023)
- [RT-1: Robotics Transformer for Real-World Control at Scale](https://mlsanity.com/paper/2212.06817/) (2022)
- [SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics](https://mlsanity.com/paper/2506.01844/) (2025)
- [Vision-Language Foundation Models as Effective Robot Imitators](https://mlsanity.com/paper/2311.01378/) (2023)
- [3D-VLA: A 3D Vision-Language-Action Generative World Model](https://mlsanity.com/paper/2403.09631/) (2024)
- [EfficientVLA: Training-Free Acceleration and Compression for Vision-Language-Action Models](https://mlsanity.com/paper/2506.10100/) (2025)
- [Discrete Diffusion VLA: Bringing Discrete Diffusion to Action Decoding in Vision-Language-Action Policies](https://mlsanity.com/paper/2508.20072/) (2025)
- [VLA-Cache: Efficient Vision-Language-Action Manipulation via Adaptive Token Caching](https://mlsanity.com/paper/2502.02175/) (2025)
- [SP-VLA: A Joint Model Scheduling and Token Pruning Approach for VLA Model Acceleration](https://mlsanity.com/paper/2506.12723/) (2025)
- [Diffusion-VLA: Generalizable and Interpretable Robot Foundation Model via Self-Generated Reasoning](https://mlsanity.com/paper/2412.03293/) (2024)
- [PyRep: Bringing V-REP to Deep Robot Learning](https://mlsanity.com/paper/1906.11176/) (2019)
- [SpecPrune-VLA: Accelerating Vision-Language-Action Models via Action-Aware Self-Speculative Pruning](https://mlsanity.com/paper/2509.05614/) (2025)
- [Running VLAs at Real-time Speed](https://mlsanity.com/paper/2510.26742/) (2025)
- [The Better You Learn, The Smarter You Prune: Towards Efficient Vision-language-action Models via Differentiable Token Pruning](https://mlsanity.com/paper/2509.12594/) (2025)
- [Think Twice, Act Once: Token-Aware Compression and Action Reuse for Efficient Inference in Vision-Language-Action Models](https://mlsanity.com/paper/2505.21200/) (2025)
- [QVLA: Not All Channels Are Equal in Vision-Language-Action Model's Quantization](https://mlsanity.com/paper/2602.03782/) (2026)
- [MLA: A Multisensory Language-Action Model for Multimodal Understanding and Forecasting in Robotic Manipulation](https://mlsanity.com/paper/2509.26642/) (2025)
- [EdgeVLA: Efficient Vision-Language-Action Models](https://mlsanity.com/paper/2507.14049/) (2025)
- [Bridging the Semantic-Action Gap in Visual Token Pruning for Efficient VLA Inference](https://mlsanity.com/paper/2511.16449/) (2025)
- [Any3D-VLA: Enhancing VLA Robustness via Diverse Point Clouds](https://mlsanity.com/paper/2602.00807/) (2026)
- [DyQ-VLA: Temporal-Dynamic-Aware Quantization for Embodied Vision-Language-Action Models](https://mlsanity.com/paper/2603.07904/) (2026)
- [KERV: Kinematic-Rectified Speculative Decoding for Embodied VLA Models](https://mlsanity.com/paper/2603.01581/) (2026)
- [Focusing on What Matters: Object-Agent-centric Tokenization for Vision Language Action models](https://mlsanity.com/paper/2509.23655/) (2025)
- [HeiSD: Hybrid Speculative Decoding for Embodied Vision-Language-Action Models with Kinematic Awareness](https://mlsanity.com/paper/2603.17573/) (2026)
- [RAPID: Redundancy-Aware and Compatibility-Optimal Edge-Cloud Partitioned Inference for Diverse VLA Models](https://mlsanity.com/paper/2603.07949/) (2026)
- [RoboECC: Multi-Factor-Aware Edge-Cloud Collaborative Deployment for VLA Models](https://mlsanity.com/paper/2603.20711/) (2026)
- [VLN-Cache: Enabling Token Caching for VLN Models with Visual/Semantic Dynamics Awareness](https://mlsanity.com/paper/2603.07080/) (2026)

## Abstract (arXiv)

Vision-Language-Action (VLA) models have emerged as the mainstream of embodied intelligence. Recent VLA models have expanded their input modalities from 2D-only to 2D+3D paradigms, forming multi-visual-modal VLA (MVLA) models. Despite achieving improved spatial perception, MVLA faces a greater acceleration demand due to the increased number of input tokens caused by modal expansion. Token pruning is an effective optimization methods tailored to MVLA models. However, existing token pruning schemes are designed for 2D-only VLA models, ignoring 2D/3D modality salience differences. In this paper, we follow the application process of multi-modal data in MVLA models and develop a tri-stage analysis to capture the discrepancy and dynamics of 2D/3D modality salience. Based on these, we propose a corresponding tri-stage token pruning framework for MVLA models to achieve optimal 2D/3D token selection and efficient pruning. Experiments show that our framework achieves up to a 2.55x inference speedup with minimal accuracy loss, while only costing 5.8% overhead. Our Code is coming soon.