Voyage Rerank 3

Voyage AI's quality-focused preview reranker, billed by all query and candidate tokens processed in each request.

Plain-English overview

What Voyage Rerank 3 actually is

Voyage Rerank 3 reorders candidates after an initial embedding or keyword search. Its processed-token price is easy to compare at a fixed workload, but the query is effectively counted once for every candidate, so cost grows with both list size and document length.

The model is currently a preview. Voyage publishes its 32K context and token rate, while several practical limits and language details on the same documentation page still describe the older 2.5 generation rather than Rerank 3.

Good fit for

Teams already using Voyage embeddings
Token-metered retrieval pipelines
Evaluating a new quality-focused reranker

Category comparison

The facts that matter for reranking models

These are provider-published specifications, not Cody benchmark scores. Follow the linked sources for current limits and endpoint-specific exceptions.

Price basis: $0.05 / 1M processed tokensThe provider's billing unit, with the document or token assumptions needed to compare it fairly.
Text capacity: 32,000 tokensHow much query and candidate text the model can consider in one scoring pass.
Candidates per request: Not published for the Rerank 3 previewThe maximum number of possible results accepted in one request, when the provider publishes it.
Language coverage: Not yet detailed for the previewPublished multilingual support; an exact count is shown only when the provider gives one.
Structured data: Text candidates; structured-data support not statedWhether the model is documented for JSON, tables, XML, or other data beyond normal prose.
Where it runs: Voyage hosted API and SDKsHosted API, cloud marketplace, private deployment, or self-hosted weights.

API and provider access

Where to get Voyage Rerank 3

Voyage AIVoyage reranker documentationCurrent model choices, API route, and context length.Voyage AIVoyage pricingProcessed-token formula and current Rerank 3 rate.

Availability

Regions and access stage

Available in preview through Voyage AI's hosted Rerank API and SDKs.

The public Rerank 3 documentation does not enumerate processing regions or a country-availability list.

Check live availability

Data and training

The route matters.

This is a concise reading of the cited provider material, not legal advice. A third-party gateway can have different storage, routing, training, and residency terms from the model maker's direct API.

Read the provider policy

Frequently asked questions

Voyage Rerank 3 FAQ

What is Voyage Rerank 3?

Voyage AI's quality-focused preview reranker, billed by all query and candidate tokens processed in each request. Voyage Rerank 3 reorders candidates after an initial embedding or keyword search. Its processed-token price is easy to compare at a fixed workload, but the query is effectively counted once for every candidate, so cost grows with both list size and document length.

When was Voyage Rerank 3 released?

The provider does not publish a clear release date for Voyage Rerank 3 in the source material reviewed by Cody.

Where can I access Voyage Rerank 3?

Available in preview through Voyage AI's hosted Rerank API and SDKs. The access routes listed in this guide are Voyage AI.

How much does Voyage Rerank 3 cost?

$0.05 / 1M processed tokens. Voyage estimates $0.0025 for a 100-document request when the query plus each document averages 500 tokens. The pricing table lists a 200M-token starter allowance.

Where is Voyage Rerank 3 available?

Available in preview through Voyage AI's hosted Rerank API and SDKs. The public Rerank 3 documentation does not enumerate processing regions or a country-availability list.

Is my Voyage Rerank 3 API data used for training?

The reviewed Rerank 3 and pricing documentation does not publish a complete prompt-retention or model-training commitment. Obtain current data-processing terms before sending private retrieval candidates. The policy belongs to the provider route and account terms, so verify it again before production use.

Related comparisons

Head-to-head comparisons

View category →