Benchmark
Agentset Embedding Recall@10 - 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
- 14
- Scores recorded
- 14
- Sources
- 1
- IQ weight
- 1.02
Score distribution
- 76.2%
- Mean
- 78.3%
- Median
- 5.8%
- Std dev
- 68.2%
- 10th percentile
- 83.9%
- 90th percentile
- 66.7%
- Min
This benchmark has a weight of 1.02 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 | voyage-4FiQa· Agentset | 84% |
| 2 | Qwen3 Embedding 4BFiQa· Agentset | 83.9% |
| 3 | Qwen3 Embedding 8BFiQa· Agentset | 81.8% |
| 4 | Gemini Embedding 2FiQa· Agentset | 81.6% |
| 5 | voyage-3.5-liteFiQa· Agentset | 79.6% |
| 6 | Text Embedding 3 LargeFiQa· Agentset | 78.3% |
| 7 | voyage-3.5FiQa· Agentset | 78.3% |
| 8 | Jina Embeddings v5 Text SmallFiQa· Agentset | 77.1% |
| 9 | Text Embedding 004FiQa· Agentset | 75.2% |
| 10 | Jina Embeddings v3FiQa· Agentset | 74.5% |
| 11 | Embed Multilingual V3.0FiQa· Agentset | 69.6% |
| 12 | Qwen3 Embedding 0.6BFiQa· Agentset | 68.3% |
| 13 | Text Embedding 3 SmallFiQa· Agentset | 68.2% |
| 14 | BGE M3FiQa· Agentset | 66.7% |