# Voyage AI Rerank 2.5

First reranker with instruction-following capabilities enabling dynamic query steering and domain-specific optimization. Supports 32K token context (8x Cohere's) through advanced distillation techniques. If you want to compare the best rerankers for your data, try [Agentset](https://app.agentset.ai/).

### Leaderboard Rank

ELO Rating

1544

Win Rate

58.0%

Accuracy (nDCG@10)

0.110

Latency

613ms

### Model Information

- **Provider**: Voyage AI  
- **License**: Proprietary  
- **Price per 1M tokens**: $0.050  
- **Release Date**: 2025-08-11  
- **Model Name**: voyage-rerank-2.5  
- **Total Evaluations**: 3300

### Performance Record

- **Wins**: 1915 (58.0%)  
- **Losses**: 1270 (38.5%)  
- **Ties**: 115 (3.5%)

## Rerankers Are Just One Piece of RAG

Agentset gives you a managed RAG pipeline with the top-ranked models and best practices baked in. No infrastructure to maintain, no reranking to configure.

Trusted by teams building production RAG applications

- **5M+ Documents**  
- **1,500+ Teams**  
- **99.9% Uptime**

## Performance Overview

Voyage AI Rerank 2.5's ELO performance varies across different benchmark datasets, showing its strengths in specific domains.

#### Voyage AI Rerank 2.5 - ELO by Dataset

| Dataset | ELO | Win Rate |  
| ------- | --- | -------- |  
| DBPedia | 1353 | 14.3% |  
| business reports | 1401 | 55.2% |  
| FiQa | 1478 | 70.1% |  
| arguana | 1508 | 90.1% |  
| MSMARCO | 1451 | 46.4% |  
| PG | 1360 | 55.5% |

## Detailed Metrics

### Dataset breakdown

Performance metrics across different benchmark datasets, including accuracy and latency percentiles.

#### DBPedia

- **ELO**: 1680  
- **Win Rate**: 58.7% (323W-188L-39T)

##### Accuracy Metrics

- nDCG@50: 0.000  
- nDCG@100: 0.000  
- Recall@50: 0.000  
- Recall@100: 0.000

##### Latency Distribution

- Mean: 583ms  
- P50 (Median): 613ms  
- P90: 632ms

#### business reports

- **ELO**: 1638  
- **Win Rate**: 61.5% (338W-202L-10T)

##### Accuracy Metrics

- nDCG@50: 0.000  
- nDCG@100: 0.000  
- Recall@50: 0.000  
- Recall@100: 0.000

##### Latency Distribution

- Mean: 612ms  
- P50 (Median): 521ms  
- P90: 734ms

#### FiQa

- **ELO**: 1627  
- **Win Rate**: 67.1% (369W-167L-14T)

##### Accuracy Metrics

- nDCG@50: 0.108  
- nDCG@100: 0.119  
- Recall@50: 0.098  
- Recall@100: 0.128

##### Latency Distribution

- Mean: 627ms  
- P50 (Median): 611ms  
- P90: 814ms

#### arguana

- **ELO**: 1508  
- **Win Rate**: 59.1% (325W-221L-4T)

##### Accuracy Metrics

- nDCG@50: 0.536  
- nDCG@100: 0.543  
- Recall@50: 0.960  
- Recall@100: 0.980

##### Latency Distribution

- Mean: 675ms  
- P50 (Median): 612ms  
- P90: 820ms

#### MSMARCO

- **ELO**: 1451  
- **Win Rate**: 46.4% (255W-247L-48T)

##### Accuracy Metrics

- nDCG@50: 0.000  
- nDCG@100: 0.000  
- Recall@50: 0.000  
- Recall@100: 0.000

##### Latency Distribution

- Mean: 571ms  
- P50 (Median): 611ms  
- P90: 647ms

#### PG

- **ELO**: 1360  
- **Win Rate**: 55.5% (305W-245L-0T)

##### Accuracy Metrics

- nDCG@50: 0.000  
- nDCG@100: 0.000  
- Recall@50: 0.000  
- Recall@100: 0.000

##### Latency Distribution

- Mean: 612ms  
- P50 (Median): 612ms  
- P90: 791ms

## Build RAG in Minutes, Not Months

Agentset gives you a complete RAG API with top-ranked rerankers and embedding models built in. Upload your data, call the API, and get accurate results from day one.

### Example Code

```typescript
import { Agentset } from "agentset";

const agentset = new Agentset();
const ns = agentset.namespace("ns_1234");

const results = await ns.search(
  "What is multi-head attention?"
);

for (const result of results) {
  console.log(result.text);
}
```

## Compare Models

### See how it stacks up

Compare Voyage AI Rerank 2.5 with other top rerankers to understand the differences in performance, accuracy, and latency.