Benchmark
MTEB(fas, v2) - Retrieval
Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages.
- Models measured
- 6
- Scores recorded
- 6
- Sources
- 1
- IQ weight
- 0.00
Score distribution
- 52.2%
- Mean
- 56%
- Median
- 7.4%
- Std dev
- 52%
- 10th percentile
- 58.9%
- 90th percentile
- 36.6%
- Min
This benchmark has a weight of 0.00 for embedding models, measured over 6 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | BGE M3Retrieval· MTEB Leaderboard data | 58.9% |
| 2 | Jina Embeddings v3Retrieval· MTEB Leaderboard data | 57.5% |
| 3 | Multilingual E5 LargeRetrieval· MTEB Leaderboard data | 56% |
| 4 | E5 Mistral 7B InstructRetrieval· MTEB Leaderboard data | 52.5% |
| 5 | Qwen3 Embedding 0.6BRetrieval· MTEB Leaderboard data | 52% |
| 6 | Text Embedding 3 SmallRetrieval· MTEB Leaderboard data | 36.6% |