rerank-2.5-lite
Rerank 2.5 Lite offers instruction-following ranking in the smaller model tier. It is a separately trained release alongside rerank-2.5.
Model design
Lite retains the 32,000-token pair limit. The release evaluates multilingual retrieval as well as domain-specific tasks. Ranking instructions describe the desired evidence rather than asking the model to generate an answer.Instruction-following rerank-2.5 models
Practical use
Practical guidance. Choose a short set of ranking criteria and check whether they reorder results as intended. Conflicting instructions can make relevance hard to measure. Keep the candidate set fixed when comparing Lite with the larger model.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | API |
| Atlas lifecycle | Legacy since 29 September 2026MongoDB Voyage AI model lifecycle: Atlas scope and release history |
| Query limit | 8,000 tokensVoyage AI reranking API and earlier models |
Related models
- rerank-2.5
Voyage AI · Voyage Rerank 2.5
Reranker that follows natural-language ranking criteria.
Launch date API - rerank-3-lite
Voyage AI · Voyage Rerank 3
Smaller instruction-following reranker for interactive retrieval.
Launch date API - rerank-2-lite
Voyage AI · Voyage Rerank 2
Smaller second-generation reranker with an 8,000-token pair limit.
Launch date API