Benchmark
VN-MTEB (vie, v1) - Clustering
Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages.
- Models measured
- 8
- Scores recorded
- 8
- Sources
- 1
- IQ weight
- 1.13
Score distribution
- 41.2%
- Mean
- 45.9%
- Median
- 11.5%
- Std dev
- 24.4%
- 10th percentile
- 51.5%
- 90th percentile
- 20.4%
- Min
This benchmark has a weight of 1.13 for embedding models, measured over 8 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 | Jina Embeddings v5 Text SmallClustering· MTEB Extra Text Boards data | 52.7% |
| 2 | Multilingual E5 Large InstructClustering· MTEB Extra Text Boards data | 51.5% |
| 3 | E5 Mistral 7B InstructClustering· MTEB Extra Text Boards data | 49% |
| 4 | BGE M3Clustering· MTEB Extra Text Boards data | 45.9% |
| 5 | Multilingual E5 LargeClustering· MTEB Extra Text Boards data | 45.8% |
| 6 | BGE multilingual gemma2Clustering· MTEB Extra Text Boards data | 40.1% |
| 7 | All MiniLM L12 V2Clustering· MTEB Extra Text Boards data | 24.4% |
| 8 | All MiniLM L6 V2Clustering· MTEB Extra Text Boards data | 20.4% |