Embedding
Nemotron 3.5 Rerank
Nemotron 3.5 RerankNemotron 3.5 Rerank scores and reorders retrieved text and visual evidence, surfacing relevant context for agents while reducing noise and token use.
Model details
View repositoryNemotron 3.5 Rerank scores retrieved candidates against a query and returns them as a reranked list, ordered by relevance. It supports text and multimodal enterprise content, helping agents find useful evidence across documents and visual materials.
By surfacing stronger evidence before generation, Nemotron 3.5 Rerank helps improve retrieval precision and reduce irrelevant context and unnecessary token use. Use it in retrieval pipelines for agents, copilots, and enterprise workflows.
