Back to all embeddings

BAAI/bge-m3

Based on XLM-RoBERTa with 8,192 token context length supporting 100+ languages with multi-functionality. Uniquely performs dense, multi-vector, and sparse retrieval simultaneously achieving SOTA on MIRACL and MKQA benchmarks. If you want to compare the best embedding models for your data, try Agentset.

Leaderboard Rank
#13
of 18
ELO Rating
1480
#13
Win Rate
44.3%
#11
Accuracy (nDCG@10)
0.674
#8
Latency
34ms
#10

Model Information

Provider
BAAI
License
Open Source
Price per 1M tokens
$0.010
Dimensions
1024
Release Date
2024-01-27
Model Name
bge-m3
Total Evaluations
830

Performance Record

Wins368 (44.3%)
Losses403 (48.6%)
Ties59 (7.1%)
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

BAAI/bge-m3's ELO performance varies across different benchmark datasets, showing its strengths in specific domains.

BAAI/bge-m3 - ELO by Dataset

Detailed Metrics

Dataset breakdown

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

business reports

ELO 150051.2% WR82W-71L-7T

Accuracy Metrics

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

Latency Distribution

Mean
27ms
P50 (Median)
27ms
P90
27ms

DBPedia

ELO 150050.6% WR81W-64L-15T

Accuracy Metrics

nDCG@5
0.801
nDCG@10
0.785
Recall@5
0.061
Recall@10
0.122

Latency Distribution

Mean
21ms
P50 (Median)
21ms
P90
21ms

FiQa

ELO 150054.7% WR82W-67L-1T

Accuracy Metrics

nDCG@5
0.743
nDCG@10
0.755
Recall@5
0.608
Recall@10
0.667

Latency Distribution

Mean
22ms
P50 (Median)
22ms
P90
22ms

SciFact

ELO 150029.4% WR47W-106L-7T

Accuracy Metrics

nDCG@5
0.571
nDCG@10
0.599
Recall@5
0.645
Recall@10
0.759

Latency Distribution

Mean
37ms
P50 (Median)
37ms
P90
37ms

MSMARCO

ELO 150035.6% WR57W-79L-24T

Accuracy Metrics

nDCG@5
0.956
nDCG@10
0.941
Recall@5
0.121
Recall@10
0.219

Latency Distribution

Mean
51ms
P50 (Median)
51ms
P90
51ms

ARCD

ELO 150047.5% WR19W-16L-5T

Accuracy Metrics

nDCG@5
0.879
nDCG@10
0.879
Recall@5
0.960
Recall@10
0.960

Latency Distribution

Mean
48ms
P50 (Median)
48ms
P90
48ms

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 BAAI/bge-m3 with other top embeddings to understand the differences in performance, accuracy, and latency.