Agentic Retrieval on BRIGHT-PRO (test)
7.17Retrieval Count (#R)ReasonIR-8B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ReasonIR-8BAgent=GPT-5-mini2026.05 | 7.17 | 4.48 | 4.42 | 3.31 | |
| INSTRUCTOR-XLAgent=GPT-5-mini2026.05 | 7.14 | 4.46 | 4.35 | 3.26 | |
| GritLM (7B)Agent=GPT-5-mini2026.05 | 7.03 | 4.51 | 4.47 | 3.36 | |
| ReasonIR-8BAgent=Qwen3.52026.05 | 6.99 | 4.1 | 3.93 | 3.08 | |
| INSTRUCTOR-XLAgent=Qwen3.52026.05 | 6.93 | 4.05 | 3.88 | 3.04 | |
| GTE-7BAgent=GPT-5-mini2026.05 | 6.67 | 4.62 | 4.51 | 3.44 | |
| DIVER-4B-1020Agent=Qwen3.52026.05 | 6.57 | 4.18 | 3.96 | 3.11 | |
| GritLM (7B)Agent=Qwen3.52026.05 | 6.48 | 4.08 | 3.85 | 3.08 | |
| Qwen3-8BAgent=GPT-5-mini2026.05 | 6.27 | 4.52 | 4.49 | 3.5 | |
| Qwen3-8BAgent=Qwen3.52026.05 | 6.16 | 4.15 | 3.93 | 3.14 | |
| RTriever-4BAgent=GPT-5-mini2026.05 | 6.01 | 4.53 | 4.43 | 3.51 | |
| DIVER-4BAgent=GPT-5-mini2026.05 | 5.91 | 4.57 | 4.46 | 3.53 | |
| GTE-7BAgent=Qwen3.52026.05 | 5.83 | 4.13 | 3.95 | 3.18 | |
| BM25Agent=GPT-5-mini2026.05 | 5.73 | 4.5 | 4.42 | 3.53 | |
| DIVER-4B-1020Agent=GPT-5-mini2026.05 | 5.61 | 4.54 | 4.43 | 3.56 | |
| DIVER-4BAgent=Qwen3.52026.05 | 5.56 | 4.2 | 4.02 | 3.29 | |
| BM25Agent=Qwen3.52026.05 | 5.21 | 4.19 | 4.01 | 3.31 | |
| BGE-Reasoner-8BAgent=GPT-5-mini2026.05 | 5.1 | 4.63 | 4.43 | 3.65 | |
| RTriever-4BAgent=Qwen3.52026.05 | 4.89 | 4.26 | 4.06 | 3.38 | |
| BGE-Reasoner-8BAgent=Qwen3.52026.05 | 4.14 | 4.22 | 3.99 | 3.44 |