Multimodal large language models (MLLMs) suffer from high computational costs due to excessive visual tokens, particularly in high-resolution and video-based scenarios. Existing token reduction methods typically focus on isolated pipeline components and often neglect textual alignment, leading to performance degradation. In this paper, we propose VisionTrim, a unified framework for training-free MLLM acceleration, integrating two effective plug-and-play modules: 1) the Dominant Vision Token Selection (DVTS) module, which preserves essential visual tokens via a global-local view, and 2) the Text-Guided Vision Complement (TGVC) module, which facilitates context-aware token merging guided by textual cues. Extensive experiments across diverse image and video multimodal benchmarks demonstrate the performance superiority of our VisionTrim, advancing practical MLLM deployment in real-world applications. The code is available at: https://github.com/hanxunyu/VisionTrim.
@article{arxiv.2601.22674,
title = {VisionTrim: Unified Vision Token Compression for Training-Free MLLM Acceleration},
author = {Hanxun Yu and Wentong Li and Xuan Qu and Song Wang and Junbo Chen and Jianke Zhu},
journal= {arXiv preprint arXiv:2601.22674},
year = {2026}
}