Back to all embeddings

Voyage 3 Large

Features 32K token context length with Matryoshka learning for flexible sizing. Offers binary quantization achieving up to 200x storage cost reduction with minimal quality loss. If you want to compare the best embedding models for your data, try Agentset.

Leaderboard Rank
#6
of 18
ELO Rating
1534
#6
Win Rate
51.3%
#6
Accuracy (nDCG@10)
0.501
#17
Latency
272ms
#15

Model Information

Provider
Voyage AI
License
Proprietary
Price per 1M tokens
$0.180
Dimensions
1024
Release Date
2025-01-07
Model Name
voyage-3-large
Total Evaluations
860

Performance Record

Wins441 (51.3%)
Losses354 (41.2%)
Ties65 (7.6%)
Wins
Losses
Ties

Embedding Models 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 embeddings to manage.

Trusted by teams building production RAG applications

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

Performance Overview

ELO ratings by dataset

Voyage 3 Large's ELO performance varies across different benchmark datasets, showing its strengths in specific domains.

Voyage 3 Large - ELO by Dataset

Detailed Metrics

Dataset breakdown

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

PG

ELO 150032.0% WR16W-33L-1T

Accuracy Metrics

nDCG@5
0.000
nDCG@10
0.000
Recall@5
0.000
Recall@10
0.000

Latency Distribution

Mean
307ms
P50 (Median)
307ms
P90
307ms

business reports

ELO 150041.7% WR75W-100L-5T

Accuracy Metrics

nDCG@5
0.000
nDCG@10
0.000
Recall@5
0.000
Recall@10
0.000

Latency Distribution

Mean
309ms
P50 (Median)
309ms
P90
309ms

DBPedia

ELO 150073.3% WR132W-34L-14T

Accuracy Metrics

nDCG@5
0.801
nDCG@10
0.790
Recall@5
0.062
Recall@10
0.123

Latency Distribution

Mean
188ms
P50 (Median)
188ms
P90
188ms

FiQa

ELO 150040.0% WR20W-30L-0T

Accuracy Metrics

nDCG@5
0.000
nDCG@10
0.000
Recall@5
0.000
Recall@10
0.000

Latency Distribution

Mean
319ms
P50 (Median)
319ms
P90
319ms

SciFact

ELO 150055.0% WR99W-78L-3T

Accuracy Metrics

nDCG@5
0.766
nDCG@10
0.779
Recall@5
0.837
Recall@10
0.878

Latency Distribution

Mean
230ms
P50 (Median)
230ms
P90
230ms

MSMARCO

ELO 150041.7% WR75W-70L-35T

Accuracy Metrics

nDCG@5
0.956
nDCG@10
0.942
Recall@5
0.122
Recall@10
0.221

Latency Distribution

Mean
251ms
P50 (Median)
251ms
P90
251ms

ARCD

ELO 150060.0% WR24W-9L-7T

Accuracy Metrics

nDCG@5
0.898
nDCG@10
0.905
Recall@5
0.960
Recall@10
0.980

Latency Distribution

Mean
300ms
P50 (Median)
300ms
P90
300ms

Build RAG in Minutes, Not Months

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

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 3 Large with other top embeddings to understand the differences in performance, accuracy, and latency.