# Prune Redundancy, Preserve Essence: Vision Token Compression in VLMs via Synergistic Importance-Diversity

arXiv: https://arxiv.org/abs/2603.09480 · HTML: https://mlsanity.com/paper/2603.09480/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- Placed in Visual token pruning and efficiency, one of 16 directions in Multimodal LLMs / vision-language models.
- First posted on arXiv 10 Mar 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Visual token pruning and efficiency](https://mlsanity.com/mllm/visual-token-pruning-and-efficiency/)
- Efficient LLM inference and serving / [Visual token pruning in multimodal models](https://mlsanity.com/inference/visual-token-pruning-in-multimodal-models/)

## Built on by (most cited first, 24 of 24)

- [OccamToken: Efficient VLM Inference with Training-Free and Budget-Adaptive Token Pruning](https://mlsanity.com/paper/2605.29657/) (2026)
- [RUTA: Principled Visual Token Allocation via Rate-Utility Optimization](https://mlsanity.com/paper/2608.04132/) (2026)
- [Not All Visual Tokens Are Equally Safe to Remove:Consequence-Sensitive Visual Token Compression](https://mlsanity.com/paper/2608.09176/) (2026)
- [ERASE: Eliminating Redundant Visual Tokens via Adaptive Two-Stage Token Pruning](https://mlsanity.com/paper/2605.09982/) (2026)
- [OTT-Vid: Optimal Transport Temporal Token Compression for Video Large Language Models](https://mlsanity.com/paper/2605.11803/) (2026)
- [DiffPrune: differentiable information throttling for token pruning in vision-language models](https://mlsanity.com/paper/2608.01985/) (2026)
- [DIVE: Dynamic Iterative Visual Evidence Construction for Efficient Vision-Language Models](https://mlsanity.com/paper/2608.04496/) (2026)
- [Centering before Pruning: Lightweight Geometry Correction for Diversity-Based Visual Token Pruning in LVLMs](https://mlsanity.com/paper/2608.30263/) (2026)
- [QCPruner: Query-Conditioned Population Coverage for Visual Token Pruning](https://mlsanity.com/paper/2609.19990/) (2026)
- [ASAP: Attention Sink Anchored Pruning](https://mlsanity.com/paper/2605.22372/) (2026)
- [DCP-Prune: Ultra-Low Token Pruning with Distribution Consistency Preservation](https://mlsanity.com/paper/2606.16633/) (2026)
- [TORINO: Token Reduction via Interpretable Concept Overlap in Vision-Language Models](https://mlsanity.com/paper/2607.04593/) (2026)
- [LAST: The Last Query Token Guides Visual Token Pruning for Edge-Cloud Collaborative MLLM Inference](https://mlsanity.com/paper/2607.27952/) (2026)
- [Keep the Needle, Prune the Haystack: Defect-Preserving Token Pruning for Efficient Zero-Shot Anomaly Detection](https://mlsanity.com/paper/2608.03681/) (2026)
- [ZIPBrain: Can EEG Foundation Models Be Faster, Locally Deployable, but Accurate?](https://mlsanity.com/paper/2608.07033/) (2026)
- [PACE: A Unified Condense-and-Extract Paradigm for Fast VLM Inference](https://mlsanity.com/paper/2608.27206/) (2026)
- [TRACE: Trajectory-robust Admission with Evidence Ordering for Efficient GUI Agents](https://mlsanity.com/paper/2609.10297/) (2026)
- [From Token Importance to Conditional Removability: Rethinking Visual Token Pruning in Multimodal Large Language Models](https://mlsanity.com/paper/2609.26484/) (2026)
- [CoViST: Visual Token Compression via Composable States](https://mlsanity.com/paper/2609.33397/) (2026)
- [ACPruner: Visual Token Pruning as Biased Attention Coverage Maximization in LVLMs](https://mlsanity.com/paper/2609.34558/) (2026)
- [SPIDER: Multi-Layer Semantic Token Pruning and Adaptive Sub-Layer Skipping in Multimodal Large Language Models](https://mlsanity.com/paper/2609.34977/) (2026)
- [Question-Specific Knowledge Graphs for Efficient Visual Reasoning](https://mlsanity.com/paper/2609.35942/) (2026)
- [When 1+1 Backfires: Rethinking Relevance-Diversity Balancing for Sample-Aware Visual Token Reduction](https://mlsanity.com/paper/2610.16615/) (2026)
- [Bifocal: Training-Free Mixed Granularity Visual Token Pruning for Multimodal LLMs](https://mlsanity.com/paper/2610.18399/) (2026)

## Abstract (arXiv)

Vision-language models (VLMs) face significant computational inefficiencies caused by excessive generation of visual tokens. While prior work shows that a large fraction of visual tokens are redundant, existing compression methods struggle to balance importance preservation and information diversity. To address this, we propose PruneSID, a training-free Synergistic Importance-Diversity approach featuring a two-stage pipeline: (1) Principal Semantic Components Analysis (PSCA) for clustering tokens into semantically coherent groups, ensuring comprehensive concept coverage, and (2) Intra-group Non-Maximum Suppression (NMS) for pruning redundant tokens while preserving key representative tokens within each group. Additionally, PruneSID incorporates an information-aware dynamic compression ratio mechanism that optimizes token compression rates based on image complexity, enabling more effective average information preservation across diverse scenes. Extensive experiments demonstrate state-of-the-art performance, achieving 96.3% accuracy on LLaVA-1.5 with only 11.1% token retention, and 92.8% accuracy at extreme compression rates (5.6%) on LLaVA-NeXT, outperforming prior methods by 2.5% with 7.8 × faster prefilling speed compared to the original model. Our framework generalizes across diverse VLMs and both image and video modalities, showcasing strong cross-modal versatility. Code is available at https://github.com/ZhengyaoFang/PruneSID.