Voyage rerank-2.5

Hosted API · Voyage AI · Updated 11 Aug 2026

Voyage AI built their rerankers with a single goal: maximise retrieval precision. The current generation, rerank-2.5 and rerank-2.5-lite, doubles the context window to 32,000 tokens and adds instruction following — you can shape what counts as relevant with a natural-language prompt instead of fine-tuning. Pricing is per token, which stays cheap when your passages are short.

Available models

The 32K context window on the 2.5 models is large enough to rerank long legal, medical or financial documents without chunking them first. Instruction following is the more interesting change: a prompt such as “prefer passages that cite a statute” reshapes the ranking without any training data.

Pricing

Voyage uses token-based pricing, which is cost-effective at high volume. Check the Voyage AI website for current rates.

Quick start

Python

pip install voyageai

import voyageai

vo = voyageai.Client(api_key="YOUR_API_KEY")

query = "How do I add reranking to my RAG pipeline?"
documents = [
    "Rerankers score each query-passage pair with a cross-encoder.",
    "BM25 is a classical keyword-based retrieval method.",
    "London is the capital of the United Kingdom.",
    "Two-stage retrieval: retrieve 50 candidates, rerank to top 5.",
]

result = vo.rerank(
    query=query,
    documents=documents,
    model="rerank-2.5",
    top_k=3,
)

for r in result.results:
    print(f"{r.relevance_score:.4f}  {documents[r.index][:80]}")

REST (curl)

curl https://api.voyageai.com/v1/rerank \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "rerank-2.5",
    "query": "How do I add reranking to my RAG pipeline?",
    "documents": ["Rerankers score...", "BM25 is..."],
    "top_k": 3
  }'

In a RAG pipeline

import voyageai

vo = voyageai.Client(api_key="YOUR_API_KEY")

def rag_answer(query: str, vector_db, llm) -> str:
    # Stage 1: retrieve wide
    candidates = vector_db.search(query, top_k=50)
    # Stage 2: rerank tight
    result = vo.rerank(query=query, documents=candidates, model="rerank-2.5", top_k=5)
    top5 = [candidates[r.index] for r in result.results]
    # Stage 3: generate
    return llm.complete(f"Context:\n" + "\n\n".join(top5) + f"\n\nQ: {query}")

Pros and cons

Other models