Skip to main content
Watch: I Built an AI App Without Leaving Databricks

Voyage AI rerankers - verified

Voyage AI reranker models

Voyage AI: 8 reranker models, read from the provider's own pages on October 18, 2026, 6 of them older models still served. Voyage AI rerankers are priced per 1M tokens, from $0.02 per 1M tokens for rerank-3-lite.

Read from docs.voyageai.com. Compare every reranker

Models
8
Current
2
Older
6
Open weights
0

Billing and caveats

Voyage AI

Per-request total token cap: "cannot exceed 600K for rerank-3, rerank-3-lite, rerank-2.5, rerank-2.5-lite, rerank-2 and rerank-2-lite; and 300K for rerank-1 and rerank-lite-1" (docs). The Batch API is listed at a 33% discount compared to standard endpoints. Docs disagree on free tokens: the Rerankers section says "The first 200 million tokens for rerank-2.5, rerank-2.5-lite, rerank-2, and rerank-2-lite are free for each account", while the Older models table, which lists those same models, says "we do not offer free tokens for them." The current-model table lists 200 million free tokens for rerank-3 and rerank-3-lite.

Current models

rerank-3

rerank-3 - API - Per 1M tokens

GA
Price (provider's unit)

$0.05 per 1M tokens

“rerank-3 · $0.00005 · $0.05 · $0.0025 · 200 million”
docs.voyageai.com

Note: Billed on total processed tokens: (query tokens x number of documents) + tokens in all documents. Table lists 200 million free tokens.

Status

GA

“Voyage currently provides the following reranker models:”
docs.voyageai.com
Context

32,000 tokens

query plus any single document

“The sum of the number of tokens in the query and the number of tokens in any single document cannot exceed 32,000 for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 16,000 for rerank-2; 8,000 for rerank-2-lite and rerank-1; and 4,000 for rerank-lite-1.”
docs.voyageai.com
Documents per request

1,000

hard limit

“The number of documents cannot exceed 1,000.”
docs.voyageai.com
Languages
Not stated
Open weights

No (API only) inferred

Note: No official open weights are listed in the sources checked; the provider documents this as an API model.

Query limit

8,000 tokens

“The query can contain a maximum of 8,000 tokens for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 4,000 tokens for rerank-2, 2,000 tokens for rerank-2-lite and rerank-1, and 1,000 tokens for rerank-lite-1.”
docs.voyageai.com

Voyage's recommended reranker on 2026-10-18 ("Highest accuracy. Recommendation for most applications."). Query limit 8,000 tokens.

rerank-3-lite

rerank-3-lite - API - Per 1M tokens

GA
Price (provider's unit)

$0.02 per 1M tokens

“rerank-3-lite · $0.00002 · $0.02 · $0.001 · 200 million”
docs.voyageai.com

Note: Billed on total processed tokens: (query tokens x number of documents) + tokens in all documents. Table lists 200 million free tokens.

Status

GA

“Voyage currently provides the following reranker models:”
docs.voyageai.com
Context

32,000 tokens

query plus any single document

“The sum of the number of tokens in the query and the number of tokens in any single document cannot exceed 32,000 for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 16,000 for rerank-2; 8,000 for rerank-2-lite and rerank-1; and 4,000 for rerank-lite-1.”
docs.voyageai.com
Documents per request

1,000

hard limit

“The number of documents cannot exceed 1,000.”
docs.voyageai.com
Languages
Not stated
Open weights

No (API only) inferred

Note: No official open weights are listed in the sources checked; the provider documents this as an API model.

Query limit

8,000 tokens

“The query can contain a maximum of 8,000 tokens for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 4,000 tokens for rerank-2, 2,000 tokens for rerank-2-lite and rerank-1, and 1,000 tokens for rerank-lite-1.”
docs.voyageai.com

Described as "Fast and cost-effective model optimized for latency-sensitive application." Query limit 8,000 tokens.

Older models

Earlier generations Voyage AI still serves or aliases to a newer model, kept apart from the current ones.

rerank-2.5

rerank-2.5 - API - Per 1M tokens

Older
Price (provider's unit)

$0.05 per 1M tokens

“rerank-1, rerank-2, rerank-2.5 · $0.00005 · $0.05”
docs.voyageai.com

Note: Billed on total processed tokens: (query tokens x number of documents) + tokens in all documents.

Status

Older

“The following are our earlier models, which are still accessible from our API.”
docs.voyageai.com
Context

32,000 tokens

query plus any single document

“The sum of the number of tokens in the query and the number of tokens in any single document cannot exceed 32,000 for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 16,000 for rerank-2; 8,000 for rerank-2-lite and rerank-1; and 4,000 for rerank-lite-1.”
docs.voyageai.com
Documents per request

1,000

hard limit

“The number of documents cannot exceed 1,000.”
docs.voyageai.com
Languages

Multilingual (no count stated)

“rerank-2.5 · 32,000 · Our generalist reranker optimized for quality with instruction-following and multilingual support.”
docs.voyageai.com
Open weights

No (API only) inferred

Note: No official open weights are listed in the sources checked; the provider documents this as an API model.

Query limit

8,000 tokens

“The query can contain a maximum of 8,000 tokens for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 4,000 tokens for rerank-2, 2,000 tokens for rerank-2-lite and rerank-1, and 1,000 tokens for rerank-lite-1.”
docs.voyageai.com

Supports instructions appended or prepended to the query. Free-token wording conflicts across the pricing page.

rerank-2.5-lite

rerank-2.5-lite - API - Per 1M tokens

Older
Price (provider's unit)

$0.02 per 1M tokens

“voyage-3-lite, rerank-lite-1, rerank-2-lite, rerank-2.5-lite · $0.00002 · $0.02”
docs.voyageai.com

Note: Billed on total processed tokens: (query tokens x number of documents) + tokens in all documents.

Status

Older

“The following are our earlier models, which are still accessible from our API.”
docs.voyageai.com
Context

32,000 tokens

query plus any single document

“The sum of the number of tokens in the query and the number of tokens in any single document cannot exceed 32,000 for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 16,000 for rerank-2; 8,000 for rerank-2-lite and rerank-1; and 4,000 for rerank-lite-1.”
docs.voyageai.com
Documents per request

1,000

hard limit

“The number of documents cannot exceed 1,000.”
docs.voyageai.com
Languages

Multilingual (no count stated)

“rerank-2.5-lite · 32,000 · Our generalist reranker optimized for both latency and quality with instruction-following and multilingual support.”
docs.voyageai.com
Open weights

No (API only) inferred

Note: No official open weights are listed in the sources checked; the provider documents this as an API model.

Query limit

8,000 tokens

“The query can contain a maximum of 8,000 tokens for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 4,000 tokens for rerank-2, 2,000 tokens for rerank-2-lite and rerank-1, and 1,000 tokens for rerank-lite-1.”
docs.voyageai.com

Instruction-following lite variant of rerank-2.5.

rerank-2

rerank-2 - API - Per 1M tokens

Older
Price (provider's unit)

$0.05 per 1M tokens

“rerank-1, rerank-2, rerank-2.5 · $0.00005 · $0.05”
docs.voyageai.com

Note: Billed on total processed tokens: (query tokens x number of documents) + tokens in all documents.

Status

Older

“The following are our earlier models, which are still accessible from our API.”
docs.voyageai.com
Context

16,000 tokens

query plus any single document

“The sum of the number of tokens in the query and the number of tokens in any single document cannot exceed 32,000 for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 16,000 for rerank-2; 8,000 for rerank-2-lite and rerank-1; and 4,000 for rerank-lite-1.”
docs.voyageai.com
Documents per request

1,000

hard limit

“The number of documents cannot exceed 1,000.”
docs.voyageai.com
Languages

Multilingual (no count stated)

“rerank-2 · 16,000 · Our generalist second-generation reranker optimized for quality with multilingual support.”
docs.voyageai.com
Open weights

No (API only) inferred

Note: No official open weights are listed in the sources checked; the provider documents this as an API model.

Query limit

4,000 tokens

“The query can contain a maximum of 8,000 tokens for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 4,000 tokens for rerank-2, 2,000 tokens for rerank-2-lite and rerank-1, and 1,000 tokens for rerank-lite-1.”
docs.voyageai.com

Query limit 4,000 tokens.

rerank-2-lite

rerank-2-lite - API - Per 1M tokens

Older
Price (provider's unit)

$0.02 per 1M tokens

“voyage-3-lite, rerank-lite-1, rerank-2-lite, rerank-2.5-lite · $0.00002 · $0.02”
docs.voyageai.com

Note: Billed on total processed tokens: (query tokens x number of documents) + tokens in all documents.

Status

Older

“The following are our earlier models, which are still accessible from our API.”
docs.voyageai.com
Context

8,000 tokens

query plus any single document

“The sum of the number of tokens in the query and the number of tokens in any single document cannot exceed 32,000 for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 16,000 for rerank-2; 8,000 for rerank-2-lite and rerank-1; and 4,000 for rerank-lite-1.”
docs.voyageai.com
Documents per request

1,000

hard limit

“The number of documents cannot exceed 1,000.”
docs.voyageai.com
Languages

Multilingual (no count stated)

“rerank-2-lite · 8000 · Our generalist second-generation reranker optimized for both latency and quality with multilingual support.”
docs.voyageai.com
Open weights

No (API only) inferred

Note: No official open weights are listed in the sources checked; the provider documents this as an API model.

Query limit

2,000 tokens

“The query can contain a maximum of 8,000 tokens for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 4,000 tokens for rerank-2, 2,000 tokens for rerank-2-lite and rerank-1, and 1,000 tokens for rerank-lite-1.”
docs.voyageai.com

Query limit 2,000 tokens.

rerank-1

rerank-1 - API - Per 1M tokens

Older
Price (provider's unit)

$0.05 per 1M tokens

“rerank-1, rerank-2, rerank-2.5 · $0.00005 · $0.05”
docs.voyageai.com

Note: Billed on total processed tokens: (query tokens x number of documents) + tokens in all documents.

Status

Older

“The following are our earlier models, which are still accessible from our API.”
docs.voyageai.com
Context

8,000 tokens

query plus any single document

“The sum of the number of tokens in the query and the number of tokens in any single document cannot exceed 32,000 for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 16,000 for rerank-2; 8,000 for rerank-2-lite and rerank-1; and 4,000 for rerank-lite-1.”
docs.voyageai.com
Documents per request

1,000

hard limit

“The number of documents cannot exceed 1,000.”
docs.voyageai.com
Languages

Multilingual (no count stated)

“rerank-1 · 8000 · Our generalist first-generation reranker optimized for quality. Multilingual support.”
docs.voyageai.com
Open weights

No (API only) inferred

Note: No official open weights are listed in the sources checked; the provider documents this as an API model.

Query limit

2,000 tokens

“The query can contain a maximum of 8,000 tokens for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 4,000 tokens for rerank-2, 2,000 tokens for rerank-2-lite and rerank-1, and 1,000 tokens for rerank-lite-1.”
docs.voyageai.com

Total-token cap per request is 300K for rerank-1 (600K for newer models).

rerank-lite-1

rerank-lite-1 - API - Per 1M tokens

Older
Price (provider's unit)

$0.02 per 1M tokens

“voyage-3-lite, rerank-lite-1, rerank-2-lite, rerank-2.5-lite · $0.00002 · $0.02”
docs.voyageai.com

Note: Billed on total processed tokens: (query tokens x number of documents) + tokens in all documents.

Status

Older

“The following are our earlier models, which are still accessible from our API.”
docs.voyageai.com
Context

4,000 tokens

query plus any single document

“The sum of the number of tokens in the query and the number of tokens in any single document cannot exceed 32,000 for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 16,000 for rerank-2; 8,000 for rerank-2-lite and rerank-1; and 4,000 for rerank-lite-1.”
docs.voyageai.com
Documents per request

1,000

hard limit

“The number of documents cannot exceed 1,000.”
docs.voyageai.com
Languages
Not stated
Open weights

No (API only) inferred

Note: No official open weights are listed in the sources checked; the provider documents this as an API model.

Query limit

1,000 tokens

“The query can contain a maximum of 8,000 tokens for rerank-3, rerank-3-lite, rerank-2.5 and rerank-2.5-lite, 4,000 tokens for rerank-2, 2,000 tokens for rerank-2-lite and rerank-1, and 1,000 tokens for rerank-lite-1.”
docs.voyageai.com

No multilingual claim in its description. Query limit 1,000 tokens.

FAQ

How much does Voyage AI reranking cost?

rerank-3: $0.05 per 1M tokens. rerank-3-lite: $0.02 per 1M tokens. rerank-2.5: $0.05 per 1M tokens. rerank-2.5-lite: $0.02 per 1M tokens. rerank-2: $0.05 per 1M tokens. rerank-2-lite: $0.02 per 1M tokens. rerank-1: $0.05 per 1M tokens. rerank-lite-1: $0.02 per 1M tokens. Prices are in Voyage AI's own units as of October 18, 2026.

What is the context length of Voyage AI rerankers?

rerank-3: 32,000 tokens, query plus any single document. rerank-3-lite: 32,000 tokens, query plus any single document. rerank-2.5: 32,000 tokens, query plus any single document. rerank-2.5-lite: 32,000 tokens, query plus any single document. rerank-2: 16,000 tokens, query plus any single document. rerank-2-lite: 8,000 tokens, query plus any single document. rerank-1: 8,000 tokens, query plus any single document. rerank-lite-1: 4,000 tokens, query plus any single document.

Can I run Voyage AI rerankers myself?

No official open weights were found in the sources checked; Voyage AI documents these as hosted models.

Sources

Last verified from each provider's own docs, pricing pages and Hugging Face cards; every quote was string-matched against the page that day. Prices are shown in the provider's own unit and change without notice, so check the linked source before you commit a budget.

Newsletter

Know when a reranker API changes

One email per week. New model releases, retrieval and pricing updates and what they mean for what you build.