Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Voyage AI / Rerank

Rerank 2 lite

Cross-encoder relevance scoring for a query and a retrieved set of documents.

Reranking
availableClosed weightsReranking
Official documentation (opens in a new tab)Reviewed 13 Oct 2026
6 of 17 specification fields collected
Context window8K8,000 tokens
Query limit2,000 query tokensSee specification notes below
Input billingSee ratesVoyage API
Output billingSee ratesPublished units and conditions below

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
8,000 tokensSource for Context / input limit (opens in a new tab)

Combined query plus one document, not the complete multi-document request.

Maximum output
Not collected
Input modalities
TextSource for Input modalities (opens in a new tab)
Output modalities
Relevance scoresSource for Output modalities (opens in a new tab)
Reasoning
Not collected
Tool calling
Not collected
Structured output
Not collected
Streaming
Not collected
Batch processing
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
2024-09-30Source for Release date (opens in a new tab)

Official release announcement date.

Parameters
Not collected
Architecture
Cross-encoder jointly processing each query–document pairSource for Architecture (opens in a new tab)
Languages
Not collected
Context caching
Not collected

No model-specific cache eligibility, retained-prefix lifetime or discounted cache rate established in the reviewed documentation.

Query limit
2,000 query tokensSource for Query limit (opens in a new tab)

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Voyage API

Provider source (opens in a new tab)
API model ID
rerank-2-lite
Access
Voyage API key; account-specific usage and rate limits apply.
Regions
Contractual residency and serving-region guarantees require verification.
API endpoint URL
https://api.voyageai.com/v1/rerankSource for API endpoint URL (opens in a new tab)
Request limits
At most 1,000 documents. Truncation defaults to true. top_k controls returned results, not the number of documents processed.Source for Request limits (opens in a new tab)
Total request input
600,000 processed tokens: query tokens × document count + sum of document tokensSource for Total request input (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 13 Oct 2026
CategoryUsageRateUnit
rerankingQuery–document processing1$0.021M tokens

Rate conditions

  1. Billable tokens = query tokens × document count + the sum of document tokens. top_k only limits returned results. Current rates for older models have no free allowance; newer-model allowances are account-specific.
Pricing source (opens in a new tab)

Data & enterprise terms

The FAQ describes an opt-out from storage and future training that enables zero-day retention on Voyage-hosted model API endpoints. This is an opt-out option, not a verified account default.

Read the service policy (opens in a new tab)

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Voyage API

rerank-2-lite
Not verified
How it works
Not collected
Minimum reusable input
Not collected
Lifetime & refresh
Not collected
How to enable it
Not collected
Usage fields
Not collected
Billing rules
Not collected
Limits & data boundaries
Not collected

No verified cache rate for this endpoint. An empty rate is not a zero charge.

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Inferred for this exact native release from provider API documentation and the provider-owned public repository inventory. Published tokenizer/configuration files are not model weights. The explicitly open-weight Voyage 4 Nano is excluded; no weight artifact for this identity was identified.

Classification evidence (opens in a new tab)

Lifecycle dates

Release announced30 Sept 2024Source (opens in a new tab)

Sources reviewed