Agentic Retrieval on BRIGHT-PRO Round 2 (test)
61.35alpha-nDCGBGE-Reasoner-8B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BGE-Reasoner-8BEvaluation Agent=GPT-5-mini2026.05 | 61.35 | 4.4 | 4.26 | |
| DIVER-4BEvaluation Agent=GPT-5-mini2026.05 | 50.29 | 4.21 | 4.12 | |
| GTE-7BEvaluation Agent=GPT-5-mini2026.05 | 49.8 | 4.17 | 4.13 | |
| DIVER-4B-1020Evaluation Agent=GPT-5-mini2026.05 | 49.21 | 4.22 | 4.06 | |
| Qwen3-8BEvaluation Agent=GPT-5-mini2026.05 | 48.77 | 4.25 | 4.11 | |
| RTriever-4BEvaluation Agent=GPT-5-mini2026.05 | 48.17 | 4.17 | 4.12 | |
| BM25Evaluation Agent=GPT-5-mini2026.05 | 48.05 | 4.19 | 4.05 | |
| GritLM (7B)Evaluation Agent=GPT-5-mini2026.05 | 43.99 | 4.19 | 4.03 | |
| ReasonIR-8BEvaluation Agent=GPT-5-mini2026.05 | 41.28 | 4.06 | 3.93 | |
| INSTRUCTOR-XLEvaluation Agent=GPT-5-mini2026.05 | 40.15 | 4.09 | 3.97 |