Benchmark
MLEB - ECHR Retrieval
Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages.
- Models measured
- 18
- Scores recorded
- 18
- Sources
- 1
- IQ weight
- 0.80
Score distribution
- 23.9%
- Mean
- 21.6%
- Median
- 10.5%
- Std dev
- 13.4%
- 10th percentile
- 37.7%
- 90th percentile
- 10.4%
- Min
This benchmark has a weight of 0.80 for embedding models, measured over 18 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 | voyage-4ECHR Retrieval· MLEB data | 46.4% |
| 2 | voyage-4-largeECHR Retrieval· MLEB data | 44.1% |
| 3 | voyage-4-liteECHR Retrieval· MLEB data | 37.7% |
| 4 | Qwen3 Embedding 4BECHR Retrieval· MLEB data | 32.2% |
| 5 | Qwen3 Embedding 8BECHR Retrieval· MLEB data | 31% |
| 6 | Qwen3 Embedding 0.6BECHR Retrieval· MLEB data | 28.7% |
| 7 | Jina Embeddings v4ECHR Retrieval· MLEB data | 24.5% |
| 8 | Jina Embeddings v5 Text NanoECHR Retrieval· MLEB data | 22.6% |
| 9 | Jina Embeddings v5 Text SmallECHR Retrieval· MLEB data | 21.6% |
| 10 | Text Embedding 3 LargeECHR Retrieval· MLEB data | 20.3% |
| 11 | Text Embedding Ada 002ECHR Retrieval· MLEB data | 20% |
| 12 | Text Embedding 3 SmallECHR Retrieval· MLEB data | 17.6% |
| 13 | snowflake-arctic-embed-l-v2.0ECHR Retrieval· MLEB data | 17.5% |
| 14 | BGE M3ECHR Retrieval· MLEB data | 17% |
| 15 | Gemini Embedding 001ECHR Retrieval· MLEB data | 13.6% |
| 16 | Granite Embedding Small English R2ECHR Retrieval· MLEB data | 13.4% |
| 17 | snowflake-arctic-embed-m-v2.0ECHR Retrieval· MLEB data | 11.2% |
| 18 | Multilingual E5 Large InstructECHR Retrieval· MLEB data | 10.4% |
Data credits: MLEB (Isaacus), MIT (licence)