Benchmark
HUME(v1) - STS
Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages.
- Models measured
- 5
- Scores recorded
- 5
- Sources
- 1
- IQ weight
- 0.48
Score distribution
- 78.7%
- Mean
- 79.9%
- Median
- 3.8%
- Std dev
- 74.3%
- 10th percentile
- 83.7%
- 90th percentile
- 74.3%
- Min
This benchmark has a weight of 0.48 for embedding models, measured over 5 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 | Qwen3 Embedding 0.6BSTS· MTEB Extra Text Boards data | 83.7% |
| 2 | Multilingual E5 LargeSTS· MTEB Extra Text Boards data | 81.2% |
| 3 | E5 Mistral 7B InstructSTS· MTEB Extra Text Boards data | 79.9% |
| 4 | All MiniLM L6 V2STS· MTEB Extra Text Boards data | 74.3% |
| 5 | All Mpnet Base V2STS· MTEB Extra Text Boards data | 74.3% |