Benchmark
MLEB - Consumer Contracts QA
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
- 1.04
Score distribution
- 85.2%
- Mean
- 85.6%
- Median
- 3.6%
- Std dev
- 83.3%
- 10th percentile
- 88.4%
- 90th percentile
- 74.2%
- Min
This benchmark has a weight of 1.04 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-4Consumer Contracts QA· MLEB data | 89% |
| 2 | voyage-4-largeConsumer Contracts QA· MLEB data | 88.8% |
| 3 | voyage-4-liteConsumer Contracts QA· MLEB data | 88.4% |
| 4 | Qwen3 Embedding 8BConsumer Contracts QA· MLEB data | 88.4% |
| 5 | Qwen3 Embedding 4BConsumer Contracts QA· MLEB data | 87.9% |
| 6 | Gemini Embedding 001Consumer Contracts QA· MLEB data | 87.6% |
| 7 | Jina Embeddings v5 Text SmallConsumer Contracts QA· MLEB data | 86.6% |
| 8 | snowflake-arctic-embed-l-v2.0Consumer Contracts QA· MLEB data | 85.7% |
| 9 | Text Embedding 3 LargeConsumer Contracts QA· MLEB data | 85.6% |
| 10 | Text Embedding 3 SmallConsumer Contracts QA· MLEB data | 85.5% |
| 11 | Text Embedding Ada 002Consumer Contracts QA· MLEB data | 85.4% |
| 12 | Jina Embeddings v4Consumer Contracts QA· MLEB data | 85.2% |
| 13 | Jina Embeddings v5 Text NanoConsumer Contracts QA· MLEB data | 85% |
| 14 | snowflake-arctic-embed-m-v2.0Consumer Contracts QA· MLEB data | 84.7% |
| 15 | Qwen3 Embedding 0.6BConsumer Contracts QA· MLEB data | 84.5% |
| 16 | BGE M3Consumer Contracts QA· MLEB data | 83.3% |
| 17 | Granite Embedding Small English R2Consumer Contracts QA· MLEB data | 78.4% |
| 18 | Multilingual E5 Large InstructConsumer Contracts QA· MLEB data | 74.2% |
Data credits: MLEB (Isaacus), MIT (licence)