arXiv:2606. 11576v1 Announce Type: cross Abstract: Modern Vision-Language Models (VLMs) benefit from chain-of-thought prompting and test-time scaling, but these gains often come with prohibitive inference cost due to large visual contexts and long decoding chains.
arXiv:2606. 07861v1 Announce Type: cross Abstract: Recent vision-language models (VLMs) excel at multimodal understanding and reasoning, yet their fine-grained visual perception remains underexplored.
arXiv:2607. 12815v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting is widely used as a test-time scaling strategy for Vision-Language Models (VLMs), but it remains unclear what is extended when VLMs generate longer reasoning traces.
Chain-of-Thought (CoT) prompting is widely used as a test-time scaling strategy for Vision-Language Models (VLMs), but it remains unclear what is extended when VLMs generate longer reasoning traces. We ask whether CoT requires continued access to image tokens, or whether it mainly operates over visual information already made available earlier in the forward pass.
The paper introduces the Capability-Driven Multimodal Scaling Law, a cross-family framework that predicts vision-language model (VLM) benchmark accuracy from a low-dimensional textual capability score extracted via PCA. By training over 150 VLMs on 34 large language models across seven families, the authors demonstrate that the law accurately extrapolates transfer rates from 8B to 72B‑parameter backbones, predicts full training trajectories, and generalizes to unseen model families. The study also reveals actionable insights, such as certain textual benchmarks negatively correlating with multimodal performance and base LLMs outperforming instruction-tuned counterparts as VLM backbones due to higher absorption rates.
Existing scaling strategies for Multimodal Large Language Models (MLLMs) typically expand either model parameters or sequential inference computation, incurring substantial memory or latency overhead. More importantly, most existing methods fail to alter the rigid, fixed computation allocation between the Vision Transformer and the Large Language Model components, limiting task-specific optimization.
arXiv:2510. 19496v3 Announce Type: replace-cross Abstract: Large vision-language models (VLMs) commonly process images at native or high resolution to remain effective across tasks.
VisCache introduces a two-stage, plug‑and‑play framework for pruning visual key‑value caches in Vision Large Language Models without retraining. The first stage filters out temporally redundant keyframes, while the second stage, PruneKV, applies a parabolic layer‑wise budget and asymmetric update to selectively prune keys and fuse values, preserving essential context. Experiments show up to 2.35× speedup and significant memory savings with only 19–28% of the original cache retained, outperforming existing baselines.
arXiv:2606. 30185v1 Announce Type: new Abstract: Improving vision-language models (VLMs) on visual reasoning typically requires retraining or hand-designed prompts and tools.
The paper introduces Adaptive Visual Token Pruning (AVTP), a training‑free framework that dynamically selects pruning layers and ratios for large vision‑language models (LVLMs) when processing multiple image sequences. By analyzing visual attention distributions across different LVLM architectures, AVTP adapts token retention to image importance, enabling efficient inference without relying on attention‑based computations incompatible with FlashAttention. Experiments show significant speedups—up to 2× for Qwen3VL‑8B—while preserving or even improving accuracy on multi‑image benchmarks.
FAVE (Foveated Adaptive Visual Encoding) is a lightweight, variable‑resolution Vision Transformer that encodes user‑selected image regions at high acuity while maintaining the image’s native geometry. In controlled experiments on small‑object ImageNet crops, FAVE outperforms a fixed‑resolution ViT by 9.4 top‑1 points while using 12.7× fewer FLOPs. When added as a local branch to FastVLM, FAVE improves TextVQA by 1.60 points and GQA attribute accuracy by 1.31 points, achieving a 3.3× speedup over SmolVLM2-2.2B with only 16 extra local tokens.
Vision-language models commonly project all tokens produced by a pretrained vision encoder into a large language model. However, final-layer features can discard text, local attributes, and spatial relationships, while high-resolution inputs substantially increase context length and inference latency.