Multi-Hop Question Answering on Bamboogle (Score)
50ScoreSearch-R1
Evaluation Results
| Method | Links | |
|---|---|---|
| Search-R1Base Model=Qwen3-4B-Instruct-25072026.04 | 50 | |
| ToolForgeBase Model=Qwen3-8B2026.04 | 48 | |
| π-PlayBase Model=Qwen3-4B-Instruct-25072026.04 | 44 | |
| Dr.ZeroBase Model=Qwen3-8B2026.04 | 40 | |
| π-PlayBase Model=Qwen3-8B2026.04 | 40 | |
| Search-R1Base Model=Qwen3-4B2026.04 | 37.6 | |
| Search-R1Base Model=Qwen3-8B2026.04 | 37.6 | |
| Dr.ZeroBase Model=Qwen3-4B-Instruct-25072026.04 | 36.8 | |
| π-PlayBase Model=Qwen3-4B2026.04 | 35.2 | |
| SQLM*Base Model=Qwen3-4B-Instruct-25072026.04 | 35.2 | |
| SQLM*Base Model=Qwen3-8B2026.04 | 32 | |
| Dr.ZeroBase Model=Qwen3-4B2026.04 | 31.2 | |
| SQLM*Base Model=Qwen3-4B2026.04 | 29.6 | |
| ReActBase Model=Qwen3-8B2026.04 | 28.8 | |
| ReActBase Model=Qwen3-4B-Instruct-25072026.04 | 27.2 | |
| ReActBase Model=Qwen3-4B2026.04 | 8.8 |