List your company

Question Answering on MMLongBench-Doc Multi-page (MP)

24.18EMDocArena
14.071216.695619.3221.9444May 27, 2026

Evaluation results

MethodLinks
DocArenaPolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0524.1819.4333.12
IKEAPolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0521.0318.8930.02
StepSearchPolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0520.9617.330.5
AutoRefinePolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0520.4717.529.82
Search-R1Policy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0520.117.9228.98
DeepResearcherPolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0519.7116.2527.16
R-SearchPolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0518.916.0228.56
ZeroSearchPolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0518.7315.8928.14
ReSearchPolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0518.4115.826.22
ASearcherPolicy Model=Qwen2.5-7B-Instruct, Max Search Turns=32026.0514.4612.1821.3