List your company

Multi-hop Question Answering on NQ, TriviaQA, Bamboogle, HotpotQA, 2WikiMultihopQA, and MuSiQue

78NQ AccuracyDCI-Agent-CC
23.9237.965266.04May 3, 2026

Evaluation results

MethodLinks
DCI-Agent-CCCategory=DCI Agents, Base Model=Sonnet 4.62026.057896808882748330.7
DCI-Agent-LiteCategory=DCI Agents, Base Model=GPT-5.4 nano2026.057284727268406815.7
R1-Searcher-7BCategory=Retrieval Agents2026.0558505446402445.37
Search-R1-32BCategory=Retrieval Agents2026.0556465244503246.75.6
Verl-Tool-Search-7B-DAPOCategory=Retrieval Agents2026.0556443250321237.714.6
ASearcher-Local-14BCategory=Retrieval Agents2026.0556586258562452.3—
ZeroSearch-7BCategory=Retrieval Agents2026.052630181018417.734.6