TAAFT
Free mode
100% free
Freemium
Free Trial
Deals
Companies
Voyage AI
Models
rerank-2-lite

rerank-2-lite

By Voyage AI
Model family: Rerank
rerank-2-lite is a cross-encoder reranker designed for the second stage of two-stage retrieval systems. Given a query and a list of candidate documents from vector or lexical search, it jointly processes each query-document pair to assign precise relevance scores for reranking. Optimized for latency, it delivers quality comparable to rerank-1 at 2.5x lower cost. Supports an 8K-token combined context length per query-document pair, with up to 2,000 tokens for the query. Natively multilingual across 31+ languages, outperforming Cohere multilingual v3 by 6.24% on multilingual retrieval benchmarks across 51 datasets. Accessible via REST API and Python/TypeScript client libraries.
Text Gen 7
Released: September 30, 2024

Overview

rerank-2-lite is a lightweight multilingual cross-encoder reranker from Voyage AI. It scores and reorders candidate documents by relevance to a query, optimized for low latency while maintaining strong retrieval quality. Supports an 8K-token combined context length per query-document pair and natively handles 31+ languages.

🔍Information retrieval🔍Document search🔍Conceptual search

About Voyage AI

Voyage AI provides best-in-class embedding models and rerankers for search and retrieval over unstructured data, used to power retrieval-augmented generation (RAG) and AI applications. It offers general-purpose, domain-specific (finance, legal, code) and company-specific fine-tuned models. Founded in 2023 and based in Palo Alto, the company was acquired by MongoDB, Inc. in February 2025 and now operates as a MongoDB subsidiary.

Industry: Artificial Intelligence
Location: Palo Alto, California, US
Website: voyageai.com
View Company Profile

Other models from this family

  • Gen 7 Rerank
    Rerank 4
    By Caldera Labs
    ReRank 4 is Cohere's latest cross-encoder reranking model for enterprise search and RAG, with 32K token context, Fast and Pro tiers, and self-learning to adapt relevance while reordering retrieved documents by semantic similarity.
    🔎Search
    Text
    Released 6mo ago
  • Gen 7 Rerank
    Rerank 3.5
    By Caldera Labs
    Rerank 3.5 is a high-accuracy cross-encoder reranking model for search and RAG. It re-scores candidate passages or documents given a query, boosting precision at the top of the results and improving answer quality downstream.
    🔎Search🔍SEO content
    Text
    Released 1y ago

Related Models

  • Gen 7
    Embed 4
    Multimodal embedding model aimed at enterprise search and retrieval over multimodal data.
    🔍Multimodal search
    Text
    Released 4mo ago
  • Gen 7
    Palmyra X 004
    Palmyra X 004 (X4) is Writer’s advanced enterprise LLM, part of the Palmyra X family. It’s tuned for reliable instruction following, reasoning, and coding with strong enterprise controls, supporting long-context prompts, JSON outputs, and tool/function calling.
    📝Writing🔍SEO content💻Coding🚀Productivity
    Text
    Released 1y ago
  • Gen 5
    Mistral 7B
    Mistral 7B is an open 7-billion parameter model built for strong quality at small size. It runs fast on a single modern GPU, follows instructions well in the Instruct variant, and delivers solid coding, summarization, and chat performance for self-hosted assistants.
    💻Coding📝Writing📚Summaries🤖ChatGPT
    Text
    Released 2y ago
Last updated: June 23, 2026
0 AIs selected
Clear selection
#
Name
Task