Benchmark
Agentset Embedding nDCG@10 - FiQa
Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages.
- Models measured
- 14
- Scores recorded
- 14
- Sources
- 1
- IQ weight
- 0.77
Score distribution
- 81.2%
- Mean
- 82.5%
- Median
- 4.3%
- Std dev
- 75.5%
- 10th percentile
- 86.7%
- 90th percentile
- 73%
- Min
This benchmark has a weight of 0.77 for embedding models, measured over 14 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 8BFiQa· Agentset | 88% |
| 2 | Text Embedding 3 LargeFiQa· Agentset | 86.7% |
| 3 | voyage-4FiQa· Agentset | 85.9% |
| 4 | Qwen3 Embedding 4BFiQa· Agentset | 83.6% |
| 5 | Gemini Embedding 2FiQa· Agentset | 83.5% |
| 6 | Jina Embeddings v5 Text SmallFiQa· Agentset | 83.1% |
| 7 | voyage-3.5FiQa· Agentset | 82.5% |
| 8 | Text Embedding 3 SmallFiQa· Agentset | 81.4% |
| 9 | Embed Multilingual V3.0FiQa· Agentset | 81.2% |
| 10 | voyage-3.5-liteFiQa· Agentset | 79.6% |
| 11 | Jina Embeddings v3FiQa· Agentset | 77.5% |
| 12 | BGE M3FiQa· Agentset | 75.5% |
| 13 | Qwen3 Embedding 0.6BFiQa· Agentset | 75.5% |
| 14 | Text Embedding 004FiQa· Agentset | 73% |