List your company

Multi-hop Question Answering on HotpotQA online Google Search API (test val)

51.3Exact Match (EM)ESearch-instruct
31.64436.74741.8546.953Oct 1, 2025

Evaluation results

MethodLinks
ESearch-instructModel Scale=3B2025.1051.361.2
ESearch-baseModel Scale=7B2025.105163.2
Search-R1-baseModel Scale=7B2025.1050.861
ESearch-instructModel Scale=7B2025.1050.764.2
Research-instructModel Scale=7B2025.1049.460.8
Search-R1-instructModel Scale=7B2025.1046.457
StepSearch-instructModel Scale=7B2025.1046.256
StepSearch-baseModel Scale=7B2025.1044.655.2
ESearch-baseModel Scale=3B2025.1043.558.6
ZeroSearch-baseModel Scale=7B2025.1043.152.9
R-Search-instruct-PPOModel Scale=3B2025.1039.849.5
StepSearch-instructModel Scale=3B2025.1039.447
ZeroSearch-instructModel Scale=7B2025.1039.448.3
SSRL-instructModel Scale=7B2025.1038.846.5
Research-baseModel Scale=7B2025.1038.648.6
R-Search-instruct-GRPOModel Scale=7B2025.1037.646.8
R-Search-instruct-GRPOModel Scale=3B2025.1037.446
R-Search-instruct-PPOModel Scale=7B2025.1035.845.3
ZeroSearch-instructModel Scale=3B2025.1035.745.3
Search-R1-instructModel Scale=3B2025.103544.2
Search-R1-baseModel Scale=3B2025.1034.843.1
SSRL-instructModel Scale=3B2025.1034.642.4
StepSearch-baseModel Scale=3B2025.1034.546.4
ZeroSearch-baseModel Scale=3B2025.1032.441.4