Benchmark
Agentset Embedding Win Rate - FiQa
Embedding benchmarks measure how well models find related items, for example when matching a search to relevant results. Scores are percentages. Higher is better.
- Models measured
- 15
- Scores recorded
- 15
- Sources
- 1
- IQ weight
- 0.42
Score distribution
- 45.8%
- Mean
- 45.3%
- Median
- 12%
- Std dev
- 26.7%
- 10th percentile
- 62.2%
- 90th percentile
- 24.7%
- Min
This benchmark has a weight of 0.42 for embedding models, measured over 15 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 | Text Embedding 3 LargeFiQa· Agentset | 65.3% |
| 2 | voyage-4FiQa· Agentset | 62.2% |
| 3 | Jina Embeddings v5 Text SmallFiQa· Agentset | 61.2% |
| 4 | BGE M3FiQa· Agentset | 54.7% |
| 5 | Qwen3 Embedding 8BFiQa· Agentset | 50.7% |
| 6 | Gemini Embedding 2FiQa· Agentset | 50.6% |
| 7 | voyage-3.5FiQa· Agentset | 49.3% |
| 8 | Qwen3 Embedding 4BFiQa· Agentset | 45.3% |
| 9 | Embed Multilingual V3.0FiQa· Agentset | 44.7% |
| 10 | voyage-3.5-liteFiQa· Agentset | 44% |
| 11 | voyage-3-largeFiQa· Agentset | 40% |
| 12 | Jina Embeddings v3FiQa· Agentset | 36.7% |
| 13 | Text Embedding 3 SmallFiQa· Agentset | 31.3% |
| 14 | Text Embedding 004FiQa· Agentset | 26.7% |
| 15 | Qwen3 Embedding 0.6BFiQa· Agentset | 24.7% |