Nemotron ColEmbed V2: Top-Performing Late Interaction embedding models for Visual Document Retrieval

By Ronay Ak, Mengyao Xu, Oliver Holworthy, Benedikt Schifferer, Zhiding Yu, Yauhen Babakhin, Radek Osmulski, Jiarui Cai, Ryan Chesler, Bo Liu and 1 more. 2026

Subfields: Information Retrieval & Search, Multimodal Learning, Representation & Self-Supervised Learning

Citations in the corpus

• LITTA: Late-Interaction and Test-Time Alignment for Visually-Grounded Multimodal Retrieval (2026, unknown): “Strong late-interaction embedding models for visual document retrieval, such as Nemotron ColEmbed V2 [25], enable efficient MaxSim-style matching and fine-grained evidence alignment.”