Rerank 5 Pro
Rerank 5 Pro is Cohere's quality-focused search reranker. It compares a query with candidate documents and can return relevant passages through Context Focus.
Model behavior
The model supports multilingual text and calibrated relevance scores. Context Focus can select passages rather than sending every part of a document onward. The API defaults max_tokens_per_doc to 4096, so the advertised context size is not the amount automatically read from every candidate.Rerank 5 releaseRerank context, truncation and input guidance
How to use it well
Practical guidance. Pro is useful when the correct evidence is present in the candidate list but appears below less useful matches. Review whether reranking moves that evidence into the final answer context. When using passage selection, check that definitions, exceptions and nearby qualifications remain available. Choose thresholds from your own reviewed queries; a score is useful for ranking but does not make a document's content true.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | rerank-v5.0-pro |
| Context window | 32,768 tokens Context is shared by the query and each candidate document. Cohere Rerank specifications |
| Model scale | Approximately 2 billion parametersCohere Rerank specifications |
| Output | Relevance scores and optional passage spansRerank 5 release |
Related models
- Rerank 5 Fast
Cohere · Rerank
Low-latency search reranker with calibrated relevance scoring and Context Focus.
Launch date not confirmedAPI - Embed 5 Pro
Cohere · Embed
Quality-focused embedding model for text, images and mixed documents, sharing a vector space with Embed 5 Fast.
Launch date not confirmedAPI - Rerank 4 Pro
Cohere · Rerank
Multilingual search reranker for documents and semi-structured records, with a focus on ranking quality.
Launch date not confirmedAPI