Multi-hop Question Answering on MuSiQue online Google Search API (val test)
26.5EMESearch-base
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ESearch-baseModel Scale=7B2025.10 | 26.5 | 37.2 | |
| ESearch-instructModel Scale=7B2025.10 | 25.4 | 37.5 | |
| ESearch-baseModel Scale=3B2025.10 | 24.7 | 36.7 | |
| StepSearch-baseModel Scale=7B2025.10 | 23.2 | 32.5 | |
| StepSearch-instructModel Scale=7B2025.10 | 22.2 | 32.7 | |
| Research-instructModel Scale=7B2025.10 | 22 | 32.1 | |
| Search-R1-baseModel Scale=7B2025.10 | 21.9 | 31 | |
| ESearch-instructModel Scale=3B2025.10 | 21.1 | 31.1 | |
| ZeroSearch-baseModel Scale=7B2025.10 | 21.1 | 29.7 | |
| StepSearch-baseModel Scale=3B2025.10 | 19.6 | 29.1 | |
| Search-R1-instructModel Scale=7B2025.10 | 18.2 | 26.8 | |
| Research-baseModel Scale=7B2025.10 | 17.6 | 27.5 | |
| R-Search-instruct-PPOModel Scale=7B2025.10 | 15.8 | 24 | |
| ZeroSearch-baseModel Scale=3B2025.10 | 15.2 | 23.7 | |
| R-Search-instruct-PPOModel Scale=3B2025.10 | 15.2 | 23.4 | |
| StepSearch-instructModel Scale=3B2025.10 | 15 | 24 | |
| R-Search-instruct-GRPOModel Scale=3B2025.10 | 14.2 | 22.7 | |
| ZeroSearch-instructModel Scale=7B2025.10 | 13.6 | 22.5 | |
| R-Search-instruct-GRPOModel Scale=7B2025.10 | 13.4 | 22.5 | |
| Search-R1-instructModel Scale=3B2025.10 | 12.8 | 19.5 | |
| Search-R1-baseModel Scale=3B2025.10 | 12 | 18.4 | |
| ZeroSearch-instructModel Scale=3B2025.10 | 11.4 | 17.6 | |
| SSRL-instructModel Scale=3B2025.10 | 11.4 | 19.5 | |
| SSRL-instructModel Scale=7B2025.10 | 10.6 | 18.4 |