ModelLLMsvrerank-v3.0

Cohere Rerank v3

by Cohere · paid · Last verified 2026-03-17

Cohere Rerank v3 is a state-of-the-art neural model designed to significantly boost the relevance of search results for Retrieval-Augmented Generation (RAG) systems. It re-scores a list of candidate documents from any keyword or vector search system, identifying the most pertinent information. It supports over 100 languages and can process long documents, making it highly versatile.

https://cohere.com/rerank ↗

B—Above Average

Adoption: B+Quality: AFreshness: B+Citations: BEngagement: F

Specifications

License: Proprietary
Pricing: paid
Capabilities: High-precision document reranking, Relevance scoring for search results, Multilingual and cross-lingual search enhancement, Long-document processing (up to 4096 tokens per passage), Integration with existing retrieval systems (keyword, vector, hybrid), Improved signal-to-noise ratio for RAG systems, Contextual understanding of query and document relationships
Integrations: [object Object], [object Object], [object Object], [object Object], [object Object]
Use Cases: [object Object], [object Object], [object Object], [object Object], [object Object]
API Available: Yes
Parameters: Undisclosed
Context Window: 4096 tokens
Modalities: text
Training Cutoff: Mid 2024
Tags: reranking, search, rag, relevance, retrieval, semantic-search, neural-search, information-retrieval, cross-lingual, llm-tooling, rag-optimization
Added: 2026-03-17
Completeness: 0.95%

Index Score

61.45

Adoption

Quality

Freshness

Citations

Engagement

Need help choosing the right model?

Get Expert Guidance

Explore the full AI ecosystem on Agents as a Service