LensVLM is a new model that compresses long contexts into images and expands only the most relevant pages for efficient processing. This open source innovation supports the broader mission to advance and democratize artificial intelligence through open science.