Single-Hop Question Answering on PopQA
73.6EMSDGA-Phase
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SDGA-PhaseFramework=ZeroSearch, Model=Qwen2.5-7B, Search Depth (Smax)=52026.05 | 73.6 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-AutoFramework=ZeroSearch, Model=Qwen2.5-7B, Search Depth (Smax)=52026.05 | 71.6 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-AutoFramework=ZeroSearch, Model=Qwen2.5-3B, Search Depth (Smax)=52026.05 | 68 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-PhaseFramework=ZeroSearch, Model=Qwen2.5-3B, Search Depth (Smax)=52026.05 | 67.9 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-PhaseFramework=Search-R1, Model=Qwen2.5-7B, Search Depth (Smax)=52026.05 | 62.4 | — | — | — | — | — | — | — | — | — | — | |
| HierSearchBackbone=Qwen2.5-7B-Instruct2026.01 | 61.6 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-AutoFramework=Search-R1, Model=Qwen2.5-7B, Search Depth (Smax)=52026.05 | 60.8 | — | — | — | — | — | — | — | — | — | — | |
| ZEROSEARCH-baseBackbone=Qwen-2.5-7B, Variant=Base2025.05 | 60.4 | — | — | — | — | — | — | — | — | — | — | |
| Dep-SearchBackbone=Qwen2.5-7B-Instruct2026.01 | 60.2 | — | — | — | — | — | — | — | — | — | — | |
| O2-SearcherBackbone=Qwen2.5-7B-Instruct2026.01 | 58.2 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-HalfFramework=ZeroSearch, Model=Qwen2.5-7B2026.05 | 57.4 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-HalfFramework=Search-R1, Model=Qwen2.5-7B2026.05 | 56.2 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-AutoFramework=Search-R1, Model=Qwen2.5-3B, Search Depth (Smax)=52026.05 | 55.9 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-PhaseFramework=Search-R1, Model=Qwen2.5-3B, Search Depth (Smax)=52026.05 | 55.4 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-FullFramework=ZeroSearch, Model=Qwen2.5-7B2026.05 | 55.2 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1-baseBackbone=Qwen-2.5-7B, Variant=Base2025.05 | 54.6 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-FullFramework=Search-R1, Model=Qwen2.5-7B2026.05 | 54.2 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-PhaseFramework=Search-R1, Model=LLaMA-3.2-3B-Base, Search Depth (Smax)=52026.05 | 54.1 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-AutoFramework=Search-R1, Model=LLaMA-3.2-3B-Base, Search Depth (Smax)=52026.05 | 53.8 | — | — | — | — | — | — | — | — | — | — | |
| R1-SearcherBackbone=Qwen2.5-7B-Instruct2026.01 | 53.2 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-PhaseFramework=ZeroSearch, Model=LLaMA-3.2-3B-Base, Search Depth (Smax)=52026.05 | 53.1 | — | — | — | — | — | — | — | — | — | — | |
| SDGA-AutoFramework=ZeroSearch, Model=LLaMA-3.2-3B-Base, Search Depth (Smax)=52026.05 | 52.3 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-HalfFramework=ZeroSearch, Model=Qwen2.5-3B2026.05 | 52.3 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1Backbone=Qwen2.5-7B-Instruct2026.01 | 51.6 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1-instBackbone=Qwen-2.5-7B, Variant=Instruct2025.05 | 51.6 | — | — | — | — | — | — | — | — | — | — | |
| ZeroSearchEvaluation Protocol=Zero-shot2026.05 | 51.5 | — | — | — | — | — | — | — | — | — | — | |
| ZeroSearchBackbone=Qwen2.5-7B-Instruct, Type=Chain-based2026.07 | 51.5 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-FullFramework=Search-R1, Model=Qwen2.5-3B2026.05 | 50.9 | — | — | — | — | — | — | — | — | — | — | |
| ZeroSearchBackbone=Qwen2.5-7B-Instruct2026.01 | 50.4 | — | — | — | — | — | — | — | — | — | — | |
| SD-Search-InstructTraining Pipeline=Pure RL, External Teacher=None, Backbone=Qwen2.5-7B2026.05 | 49.4 | — | — | — | — | — | — | — | — | — | — | |
| SD-Search-BaseTraining Pipeline=Pure RL, External Teacher=None, Backbone=Qwen2.5-7B2026.05 | 48.9 | — | — | — | — | — | — | — | — | — | — | |
| HierSearchBackbone=Qwen2.5-3B-Instruct2026.01 | 48.8 | — | — | — | — | — | — | — | — | — | — | |
| ZEROSEARCH-instBackbone=Qwen-2.5-7B, Variant=Instruct2025.05 | 48.8 | — | — | — | — | — | — | — | — | — | — | |
| AutoRefineBackbone=Qwen2.5-7B-Instruct2026.03 | 48.7 | — | — | — | — | — | — | — | — | — | — | |
| AutoRefine-BaseTraining Pipeline=Pure RL, External Teacher=None, Backbone=Qwen2.5-7B2026.05 | 48.7 | — | — | — | — | — | — | — | — | — | — | |
| SKILLGRAPHEvaluation Protocol=Zero-shot2026.05 | 48.5 | — | — | — | — | — | — | — | — | — | — | |
| Thinker-InstructTraining Pipeline=Synthesize → SFT, External Teacher=Qwen2.5-72B, Backbone=Qwen2.5-7B2026.05 | 48.4 | — | — | — | — | — | — | — | — | — | — | |
| ZEROSEARCH-baseBackbone=LLaMA-3.2-3B, Variant=Base2025.05 | 48.4 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-FullFramework=ZeroSearch, Model=Qwen2.5-3B2026.05 | 48.1 | — | — | — | — | — | — | — | — | — | — | |
| IGRPOBackbone=Qwen2.5-3B-Instruct, Type=Tree-based2026.07 | 47.5 | — | — | — | — | — | — | — | — | — | — | |
| Dep-SearchBackbone=Qwen2.5-3B-Instruct2026.01 | 47.4 | — | — | — | — | — | — | — | — | — | — | |
| Tree-GRPOBackbone=Qwen2.5-14b2025.09 | 47.3 | — | — | — | — | — | — | — | — | — | — | |
| MR-Search-BaseTraining Pipeline=Pure RL, External Teacher=None, Backbone=Qwen2.5-7B2026.05 | 47.2 | — | — | — | — | — | — | — | — | — | — | |
| GSPOBackbone=Qwen2.5-14b2025.09 | 47.1 | — | — | — | — | — | — | — | — | — | — | |
| IGRPOBackbone=Qwen2.5-7B-Instruct, Type=Tree-based2026.07 | 47 | — | — | — | — | — | — | — | — | — | — | |
| MemPO2026.02 | 46.7 | — | — | 53.53 | — | — | 0.25 | 14 | — | — | — | |
| GRPOBackbone=Qwen2.5-14b2025.09 | 46.7 | — | — | — | — | — | — | — | — | — | — | |
| GRPOmemory=none2026.02 | 46.5 | — | — | 53.77 | — | — | 1 | 26 | — | — | — | |
| GRPO-FullFramework=ZeroSearch, Model=LLaMA-3.2-3B-Base2026.05 | 46.5 | — | — | — | — | — | — | — | — | — | — | |
| GSPOBackbone=Qwen2.5-7b2025.09 | 46.1 | — | — | — | — | — | — | — | — | — | — | |
| GiGPOBackbone=Qwen2.5-7B-Instruct, Type=Chain-based2026.07 | 46.1 | — | — | — | — | — | — | — | — | — | — | |
| ZEROSEARCH-instBackbone=LLaMA-3.2-3B, Variant=Instruct2025.05 | 46 | — | — | — | — | — | — | — | — | — | — | |
| SkillRLEvaluation Protocol=Zero-shot2026.05 | 45.9 | — | — | — | — | — | — | — | — | — | — | |
| MultiSearchBackbone=3B-Base2026.05 | 45.5 | — | — | 49.7 | — | — | — | — | — | 47.1 | — | |
| IGPOBackbone=Qwen2.5-7B-Instruct, Type=Chain-based2026.07 | 45.4 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-HalfFramework=Search-R1, Model=Qwen2.5-3B2026.05 | 45.3 | — | — | — | — | — | — | — | — | — | — | |
| CoDA-BaseBackbone=Qwen2.5-3B, Retrieval Protocol=w/ Multi-Hop Retrieval2025.12 | 45.2 | — | — | — | — | — | — | — | — | — | — | |
| Tree-GRPOBackbone=Qwen2.5-7b2025.09 | 45.2 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-FullFramework=Search-R1, Model=LLaMA-3.2-3B-Base2026.05 | 45.1 | — | — | — | — | — | — | — | — | — | — | |
| AutoRefineBackbone=Qwen2.5-3B-Instruct2026.03 | 45 | — | — | — | — | — | — | — | — | — | — | |
| AutoRefine2026.05 | 45 | — | — | 47.9 | — | — | — | — | — | 46.8 | — | |
| AutoRefine-BaseBackbone=Qwen2.5-3B, Retrieval Protocol=w/ Multi-Hop Retrieval2025.12 | 44.9 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1-BaseTraining Pipeline=Pure RL, External Teacher=None, Backbone=Qwen2.5-7B2026.05 | 44.9 | — | — | — | — | — | — | — | — | — | — | |
| ZEROSEARCH-instBackbone=Qwen-2.5-3B, Variant=Instruct2025.05 | 44.8 | — | — | — | — | — | — | — | — | — | — | |
| ZeroSearchBackbone=Qwen2.5-3B-Instruct, Type=Chain-based2026.07 | 44.8 | — | — | — | — | — | — | — | — | — | — | |
| EvolveREvaluation Protocol=Zero-shot2026.05 | 44.6 | — | — | — | — | — | — | — | — | — | — | |
| GRPOBackbone=Qwen2.5-7b2025.09 | 44.3 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1-baseBackbone=Qwen-2.5-3B, Variant=Base2025.05 | 44.2 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1-instBackbone=LLaMA-3.2-3B, Variant=Instruct2025.05 | 44.2 | — | — | — | — | — | — | — | — | — | — | |
| Tree-GRPOModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.02 | 44.14 | — | 1.27 | — | — | — | — | — | — | — | — | |
| Search-R1-baseBackbone=LLaMA-3.2-3B, Variant=Base2025.05 | 44 | — | — | — | — | — | — | — | — | — | — | |
| ReSearch2026.02 | 43.8 | — | — | 52.41 | — | — | 0.23 | 16 | — | — | — | |
| Tree-GRPOBackbone=Qwen2.5-3B-Instruct, Type=Tree-based2026.07 | 43.7 | — | — | — | — | — | — | — | — | — | — | |
| EvalActBackbone=Qwen2.5-7B-Instruct2026.03 | 43.6 | — | — | — | — | — | — | — | — | — | — | |
| Tree-GRPOBackbone=Qwen2.5-3b2025.09 | 43.6 | — | — | — | — | — | — | — | — | — | — | |
| SIGHTModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.02 | 43.2 | — | 1.11 | — | — | — | — | — | — | — | — | |
| EvalActBackbone=Qwen2.5-3B-Instruct2026.03 | 43.2 | — | — | — | — | — | — | — | — | — | — | |
| IGPOBackbone=Qwen2.5-3B-Instruct, Type=Chain-based2026.07 | 43.2 | — | — | — | — | — | — | — | — | — | — | |
| ReSearch-BaseBackbone=Qwen2.5-3B, Retrieval Protocol=w/ Multi-Hop Retrieval2025.12 | 43 | — | — | — | — | — | — | — | — | — | — | |
| GRPO-HalfFramework=ZeroSearch, Model=LLaMA-3.2-3B-Base2026.05 | 42.8 | — | — | — | — | — | — | — | — | — | — | |
| Tree-GRPOBackbone=Qwen2.5-7B-Instruct, Type=Tree-based2026.07 | 42.8 | — | — | — | — | — | — | — | — | — | — | |
| ARPOModel=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.02 | 42.6 | — | 1.95 | — | — | — | — | — | — | — | — | |
| GSPOBackbone=Qwen2.5-3b2025.09 | 42.5 | — | — | — | — | — | — | — | — | — | — | |
| GRPOBackbone=Llama3.2-3b2025.09 | 42.4 | — | — | — | — | — | — | — | — | — | — | |
| GiGPOBackbone=Qwen2.5-3B-Instruct, Type=Chain-based2026.07 | 42.4 | — | — | — | — | — | — | — | — | — | — | |
| Tree-GRPOBackbone=Llama3.2-3b2025.09 | 42.3 | — | — | — | — | — | — | — | — | — | — | |
| GRPOBackbone=Qwen2.5-3b2025.09 | 42 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1Model=Qwen2.5-7B-Instruct, Retrieval Configuration=w/ Retrieval2026.02 | 41.8 | — | 1.69 | — | — | — | — | — | — | — | — | |
| Tree-GRPOBackbone=Qwen2.5-1.5b2025.09 | 41.6 | — | — | — | — | — | — | — | — | — | — | |
| ZEROSEARCH-baseBackbone=Qwen-2.5-3B, Variant=Base2025.05 | 41.4 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1-BaseBackbone=Qwen2.5-3B, Retrieval Protocol=w/ Multi-Hop Retrieval2025.12 | 41.3 | — | — | — | — | — | — | — | — | — | — | |
| Search-R12026.05 | 41.3 | — | — | 42.9 | — | — | — | — | — | 44.2 | — | |
| SIGHTModel=Qwen2.5-3B-Instruct, Retrieval Configuration=w/ Retrieval2026.02 | 41.21 | — | 1.22 | — | — | — | — | — | — | — | — | |
| GRPO-HalfFramework=Search-R1, Model=LLaMA-3.2-3B-Base2026.05 | 41.2 | — | — | — | — | — | — | — | — | — | — | |
| DeepResearcher2026.02 | 40.8 | — | — | 47.9 | — | — | 0.6 | 22 | — | — | — | |
| GSPOBackbone=Llama3.2-3b2025.09 | 40.8 | — | — | — | — | — | — | — | — | — | — | |
| qwen2.5Mode=ReAct2026.02 | 40.5 | — | — | 47.77 | — | — | 1.43 | 26 | — | — | — | |
| O2-SearcherBackbone=Qwen2.5-3B-Instruct2026.01 | 40.4 | — | — | — | — | — | — | — | — | — | — | |
| Search-R1Backbone=Qwen2.5-7B-Instruct2026.03 | 39.7 | — | — | — | — | — | — | — | — | — | — | |
| GRPOBackbone=Qwen2.5-1.5b2025.09 | 39.7 | — | — | — | — | — | — | — | — | — | — |