BAAI/bge-m3 vs OpenAI text-embedding-3-large
Detailed comparison between BAAI/bge-m3 and OpenAI text-embedding-3-large. See which embedding best meets your accuracy and performance needs.
Model Comparison
OpenAI text-embedding-3-large takes the lead.
Both BAAI/bge-m3 and OpenAI text-embedding-3-large are powerful embedding models designed to improve retrieval quality in RAG applications. However, their performance characteristics differ in important ways.
Why OpenAI text-embedding-3-large:
- OpenAI text-embedding-3-large has 48 higher ELO rating
- OpenAI text-embedding-3-large delivers better accuracy (nDCG@10: 0.811 vs 0.753)
- OpenAI text-embedding-3-large has a 14.8% higher win rate
Overview
Key metrics
ELO Rating
Overall ranking quality
BAAI/bge-m3
OpenAI text-embedding-3-large
Win Rate
Head-to-head performance
BAAI/bge-m3
OpenAI text-embedding-3-large
Accuracy (nDCG@10)
Ranking quality metric
BAAI/bge-m3
OpenAI text-embedding-3-large
Average Latency
Response time
BAAI/bge-m3
OpenAI text-embedding-3-large
Visual Performance Analysis
Performance
ELO Rating Comparison
Win/Loss/Tie Breakdown
Accuracy Across Datasets (nDCG@10)
Latency Distribution (ms)
Breakdown
How the models stack up
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Overall Performance | |||
| ELO Rating | 1491 | 1539 | Overall ranking quality based on pairwise comparisons |
| Win Rate | 40.9% | 55.7% | Percentage of comparisons won against other models |
| Pricing & Availability | |||
| Price per 1M tokens | $0.010 | $0.130 | Cost per million tokens processed |
| Dimensions | 1024 | 3072 | Vector embedding dimensions (lower is more efficient) |
| Release Date | 2024-01-27 | 2024-01-25 | Model release date |
| Accuracy Metrics | |||
| Avg nDCG@10 | 0.753 | 0.811 | Normalized discounted cumulative gain at position 10 |
| Performance Metrics | |||
| Avg Latency | 29ms | 11ms | Average response time across all datasets |
Dataset Performance
By field
Comprehensive comparison of accuracy metrics (nDCG, Recall) and latency percentiles for each benchmark dataset.
PG
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Accuracy Metrics | |||
| Latency Metrics | |||
| Mean | 16ms | 10ms | Average response time |
| P50 | 16ms | 10ms | 50th percentile (median) |
| P90 | 19ms | 11ms | 90th percentile |
business reports
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Accuracy Metrics | |||
| Latency Metrics | |||
| Mean | 34ms | 10ms | Average response time |
| P50 | 33ms | 10ms | 50th percentile (median) |
| P90 | 39ms | 12ms | 90th percentile |
DBPedia
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Accuracy Metrics | |||
| nDCG@5 | 0.625 | 0.648 | Ranking quality at top 5 results |
| nDCG@10 | 0.603 | 0.641 | Ranking quality at top 10 results |
| Recall@5 | 0.236 | 0.255 | % of relevant docs in top 5 |
| Recall@10 | 0.341 | 0.377 | % of relevant docs in top 10 |
| Latency Metrics | |||
| Mean | 17ms | 8ms | Average response time |
| P50 | 17ms | 8ms | 50th percentile (median) |
| P90 | 20ms | 9ms | 90th percentile |
FiQa
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Accuracy Metrics | |||
| nDCG@5 | 0.597 | 0.730 | Ranking quality at top 5 results |
| nDCG@10 | 0.609 | 0.752 | Ranking quality at top 10 results |
| Recall@5 | 0.607 | 0.700 | % of relevant docs in top 5 |
| Recall@10 | 0.666 | 0.781 | % of relevant docs in top 10 |
| Latency Metrics | |||
| Mean | 32ms | 10ms | Average response time |
| P50 | 31ms | 9ms | 50th percentile (median) |
| P90 | 37ms | 11ms | 90th percentile |
SciFact
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Accuracy Metrics | |||
| nDCG@5 | 0.578 | 0.726 | Ranking quality at top 5 results |
| nDCG@10 | 0.617 | 0.761 | Ranking quality at top 10 results |
| Recall@5 | 0.652 | 0.768 | % of relevant docs in top 5 |
| Recall@10 | 0.763 | 0.863 | % of relevant docs in top 10 |
| Latency Metrics | |||
| Mean | 31ms | 11ms | Average response time |
| P50 | 30ms | 11ms | 50th percentile (median) |
| P90 | 35ms | 13ms | 90th percentile |
MSMARCO
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Accuracy Metrics | |||
| nDCG@5 | 0.997 | 1.000 | Ranking quality at top 5 results |
| nDCG@10 | 0.997 | 1.000 | Ranking quality at top 10 results |
| Recall@5 | 0.122 | 0.123 | % of relevant docs in top 5 |
| Recall@10 | 0.220 | 0.224 | % of relevant docs in top 10 |
| Latency Metrics | |||
| Mean | 19ms | 8ms | Average response time |
| P50 | 18ms | 8ms | 50th percentile (median) |
| P90 | 22ms | 9ms | 90th percentile |
NorQuAD
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Accuracy Metrics | |||
| Latency Metrics | |||
| Mean | 41ms | 14ms | Average response time |
| P50 | 40ms | 14ms | 50th percentile (median) |
| P90 | 47ms | 16ms | 90th percentile |
ARCD
| Metric | BAAI/bge-m3 | OpenAI text-embedding-3-large | Description |
|---|---|---|---|
| Accuracy Metrics | |||
| nDCG@5 | 0.941 | 0.899 | Ranking quality at top 5 results |
| nDCG@10 | 0.941 | 0.899 | Ranking quality at top 10 results |
| Recall@5 | 0.960 | 0.940 | % of relevant docs in top 5 |
| Recall@10 | 0.960 | 0.940 | % of relevant docs in top 10 |
| Latency Metrics | |||
| Mean | 40ms | 14ms | Average response time |
| P50 | 40ms | 14ms | 50th percentile (median) |
| P90 | 47ms | 16ms | 90th percentile |
Explore More
Compare more embeddings
See how all embedding models stack up. Compare OpenAI, Cohere, Jina AI, Voyage, and more. View comprehensive benchmarks, compare performance metrics, and find the perfect embedding for your RAG application.