Agentic Search on Bamboogle (String-F1)
73.1String-F1 ScoreVanilla-Qwen3-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| Vanilla-Qwen3-32BASP=false2026.04 | 73.1 | |
| Vanilla-Qwen3-8BASP=false2026.04 | 71.5 | |
| SFT-Qwen3-1.7BASP=true2026.04 | 70.6 | |
| Mixed-Qwen3-1.7BASP=true2026.04 | 69.4 | |
| Vanilla-Qwen3-4BASP=false2026.04 | 69.1 | |
| SFT-Llama-3.2-3BASP=true2026.04 | 68.4 | |
| SFT-Llama-3.2-1BASP=true2026.04 | 64.6 | |
| SFT-Qwen3-0.6BASP=true2026.04 | 62.9 | |
| OPD-Qwen3-1.7BASP=true2026.04 | 61.4 | |
| Distilled-Llama3.2-3BASP=false2026.04 | 53.6 | |
| Distilled-Qwen3-1.7BASP=false2026.04 | 53.2 | |
| Vanilla-Qwen3-1.7BASP=false2026.04 | 50.6 | |
| Distilled-Llama3.2-1BASP=false2026.04 | 37.9 | |
| Vanilla-Qwen3-0.6BASP=false2026.04 | 34.4 |