List your company

Single-Hop Question Answering on NQ (in-domain)

48.2AccuracyGEPO
10.13620.01829.939.782Oct 30, 2025Nov 16, 2025Dec 3, 2025Dec 20, 2025Jan 6, 2026Jan 23, 2026Feb 9, 2026

Evaluation results

MethodLinks
GEPOType=RL Training, Model=Qwen2.5-7B-Instruct2025.1048.2
GiGPOType=RL Training, Model=Qwen2.5-7B-Instruct2025.1046.4
SKILLRLBackbone=Qwen2.5-7B-Instruct2026.0245.9
ZeroSearchBackbone=Qwen2.5-7B-Instruct2026.0243.6
GEPOType=RL Training, Model=Qwen2.5-3B-Instruct2025.1043.6
ZeroSearchType=RL Training, Model=Qwen2.5-7B-Instruct2025.1043.6
EvolveRBackbone=Qwen2.5-7B-Instruct2026.0243.5
GiGPOType=RL Training, Model=Qwen2.5-3B-Instruct2025.1042
ZeroSearchType=RL Training, Model=Qwen2.5-3B-Instruct2025.1041.4
Search-R1Backbone=Qwen2.5-7B-Instruct2026.0239.3
Search-R1Type=RL Training, Model=Qwen2.5-7B-Instruct2025.1039.3
Search-R1Type=RL Training, Model=Qwen2.5-3B-Instruct2025.1034.1
RAGBackbone=Qwen2.5-7B-Instruct2026.0227.4
R1-InstructType=RL Training, Model=Qwen2.5-3B-Instruct2025.1027
R1-InstructBackbone=Qwen2.5-7B-Instruct2026.0221
R1-InstructType=RL Training, Model=Qwen2.5-7B-Instruct2025.1021
Search-o1Backbone=Qwen2.5-7B-Instruct2026.0219.4
CoTBackbone=Qwen2.5-7B-Instruct2026.0212.8
Qwen2.5Backbone=Qwen2.5-7B-Instruct2026.0211.6