# BAAI/bge-m3 vs Qwen3 Embedding 0.6B

Detailed comparison between BAAI/bge-m3 and Qwen3 Embedding 0.6B. See which embedding best meets your accuracy and performance needs. If you want to compare these models on your data, try [Agentset](https://app.agentset.ai/).

## Model Comparison

## BAAI/bge-m3 takes the lead.

Both BAAI/bge-m3 and Qwen3 Embedding 0.6B are powerful embedding models designed to improve retrieval quality in RAG applications. However, their performance characteristics differ in important ways.

### Why BAAI/bge-m3:

- BAAI/bge-m3 has 60 higher ELO rating
- BAAI/bge-m3 delivers better accuracy (nDCG@10: 0.674 vs 0.656)
- BAAI/bge-m3 has a 8.7% higher win rate

## Overview

## Key metrics

### ELO Rating

Overall ranking quality

| Model | ELO Rating |
| --- | --- |
| BAAI/bge-m3 | 1480 |
| Qwen3 Embedding 0.6B | 1420 |

### Win Rate

Head-to-head performance

| Model | Win Rate |
| --- | --- |
| BAAI/bge-m3 | 44.3% |
| Qwen3 Embedding 0.6B | 35.7% |

### Accuracy (nDCG@10)

Ranking quality metric

| Model | nDCG@10 |
| --- | --- |
| BAAI/bge-m3 | 0.674 |
| Qwen3 Embedding 0.6B | 0.656 |

### Average Latency

Response time

| Model | Average Latency |
| --- | --- |
| BAAI/bge-m3 | 34ms |
| Qwen3 Embedding 0.6B | 25ms |

## Visual Performance Analysis

## Performance

#### ELO Rating Comparison

Qwen3 Embedding 0.6B  BAAI/bge-m3  4000 8000 12000 15000

#### Win/Loss/Tie Breakdown

- BAAI/bge-m3
- Qwen3 Embedding 0.6B

#### Accuracy Across Datasets (nDCG@10)

- BAAI/bge-m3
- Qwen3 Embedding 0.6B

DBPedia FiQa SciFact MSMARCO ARCD  0.25 0.50 0.75 nDCG@10

#### Latency Distribution (ms)

- BAAI/bge-m3
- Qwen3 Embedding 0.6B

Mean P50 P90  918 2736 Latency (ms)

## Breakdown

## How the models stack up

| Metric | BAAI/bge-m3 | Qwen3 Embedding 0.6B | Description |
| --- | --- | --- | --- |
| Overall Performance | | | |
| ELO Rating | 1480 | 1420 | Overall ranking quality based on pairwise comparisons |
| Win Rate | 44.3% | 35.7% | Percentage of comparisons won against other models |
| Pricing & Availability | | | |
| Price per 1M tokens | $0.010 | $0.010 | Cost per million tokens processed |
| Dimensions | 1024 | 1024 | Vector embedding dimensions (lower is more efficient) |
| Release Date | 2024-01-27 | 2025-06-06 | Model release date |
| Accuracy Metrics | | | |
| Avg nDCG@10 | 0.674 | 0.656 | Normalized discounted cumulative gain at position 10 |
| Performance Metrics | | | |
| Avg Latency | 34ms | 25ms | Average response time across all datasets |

## Dataset Performance

## By field

Comprehensive comparison of accuracy metrics (nDCG, Recall) and latency percentiles for each benchmark dataset.

### business reports

| Metric | BAAI/bge-m3 | Qwen3 Embedding 0.6B | Description |
| --- | --- | --- | --- |
| Accuracy Metrics | | | |
| nDCG@5 | 0.000 | 0.000 | Ranking quality at top 5 results |
| nDCG@10 | 0.000 | 0.000 | Ranking quality at top 10 results |
| Recall@5 | 0.000 | 0.000 | % of relevant docs in top 5 |
| Recall@10 | 0.000 | 0.000 | % of relevant docs in top 10 |
| Latency Metrics | | | |
| Mean | 27ms | 21ms | Average response time |
| P50 | 27ms | 21ms | 50th percentile (median) |
| P90 | 27ms | 21ms | 90th percentile |

### DBPedia

| Metric | BAAI/bge-m3 | Qwen3 Embedding 0.6B | Description |
| --- | --- | --- | --- |
| Accuracy Metrics | | | |
| nDCG@5 | 0.801 | 0.716 | Ranking quality at top 5 results |
| nDCG@10 | 0.785 | 0.730 | Ranking quality at top 10 results |
| Recall@5 | 0.061 | 0.053 | % of relevant docs in top 5 |
| Recall@10 | 0.122 | 0.105 | % of relevant docs in top 10 |
| Latency Metrics | | | |
| Mean | 21ms | 13ms | Average response time |
| P50 | 21ms | 13ms | 50th percentile (median) |
| P90 | 21ms | 13ms | 90th percentile |

### FiQa

| Metric | BAAI/bge-m3 | Qwen3 Embedding 0.6B | Description |
| --- | --- | --- | --- |
| Accuracy Metrics | | | |
| nDCG@5 | 0.743 | 0.755 | Ranking quality at top 5 results |
| nDCG@10 | 0.755 | 0.755 | Ranking quality at top 10 results |
| Recall@5 | 0.608 | 0.591 | % of relevant docs in top 5 |
| Recall@10 | 0.667 | 0.683 | % of relevant docs in top 10 |
| Latency Metrics | | | |
| Mean | 22ms | 19ms | Average response time |
| P50 | 22ms | 19ms | 50th percentile (median) |
| P90 | 22ms | 19ms | 90th percentile |

### SciFact

| Metric | BAAI/bge-m3 | Qwen3 Embedding 0.6B | Description |
| --- | --- | --- | --- |
| Accuracy Metrics | | | |
| nDCG@5 | 0.571 | 0.658 | Ranking quality at top 5 results |
| nDCG@10 | 0.599 | 0.666 | Ranking quality at top 10 results |
| Recall@5 | 0.645 | 0.718 | % of relevant docs in top 5 |
| Recall@10 | 0.759 | 0.779 | % of relevant docs in top 10 |
| Latency Metrics | | | |
| Mean | 37ms | 62ms | Average response time |
| P50 | 37ms | 62ms | 50th percentile (median) |
| P90 | 37ms | 62ms | 90th percentile |

### MSMARCO

| Metric | BAAI/bge-m3 | Qwen3 Embedding 0.6B | Description |
| --- | --- | --- | --- |
| Accuracy Metrics | | | |
| nDCG@5 | 0.956 | 0.943 | Ranking quality at top 5 results |
| nDCG@10 | 0.941 | 0.933 | Ranking quality at top 10 results |
| Recall@5 | 0.121 | 0.122 | % of relevant docs in top 5 |
| Recall@10 | 0.219 | 0.215 | % of relevant docs in top 10 |
| Latency Metrics | | | |
| Mean | 51ms | 15ms | Average response time |
| P50 | 51ms | 15ms | 50th percentile (median) |
| P90 | 51ms | 15ms | 90th percentile |

### ARCD

| Metric | BAAI/bge-m3 | Qwen3 Embedding 0.6B | Description |
| --- | --- | --- | --- |
| Accuracy Metrics | | | |
| nDCG@5 | 0.879 | 0.757 | Ranking quality at top 5 results |
| nDCG@10 | 0.879 | 0.763 | Ranking quality at top 10 results |
| Recall@5 | 0.960 | 0.880 | % of relevant docs in top 5 |
| Recall@10 | 0.960 | 0.900 | % of relevant docs in top 10 |
| Latency Metrics | | | |
| Mean | 48ms | 18ms | Average response time |
| P50 | 48ms | 18ms | 50th percentile (median) |
| P90 | 48ms | 18ms | 90th percentile |

## Code Example

### TypeScript

```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);
}
```

### Python

```python
# Python example to query the models
# Add your implementation here
```