Apple's LensVLM-9B shrinks long documents into images, then expands only the pages that matter
LensVLM: Compressing long context as images, expanding only relevant pages
Apple released LensVLM-9B, a 9B vision language model that compresses long text into images and selectively expands only relevant pages to their uncompressed form using learned tools. Built on Qwen3.5-9B, it offers 5x, 10x, and 15x compression options. The model is available on Hugging Face under Apple's ML research license, with code on GitHub and a paper on arXiv.
LensVLM is a 9B Vision Language Model (VLM) that scans compressed images of text, then selectively expands only the relevant pages to their uncompressed form via learned tools.