List your company

Multi-hop Retrieval on Average (HotpotQA, 2WikiMultihopQA, Musique, Morehopqa) (test)

78.44Average RecallThinkGR
41.457651.058860.6670.2612May 21, 2026

Evaluation results

MethodLinks
ThinkGRModel Parameters=8B2026.0578.44
w/o ThoughtModel Parameters=8B2026.0573.53
w/o RLModel Parameters=8B2026.0571.73
GritHopperModel Parameters=7B2026.0571.58
R3-RAGModel Parameters=8B2026.0565.88
RT-RAGModel Parameters=8B2026.0561.47
IRCoTModel Parameters=70B2026.0559.42
ITER-RETGENModel Parameters=70B2026.0555.5
Auto-RAGModel Parameters=7B2026.0555.17
MDRModel Parameters=110M2026.0554.73
BGE-largeModel Parameters=326M2026.0549.97
SEALModel Parameters=406M2026.0546.84
w/o SFTModel Parameters=8B2026.0546.15
SelfaskModel Parameters=70B2026.0545.84
Contriever2026.0545.17
BM252026.0542.88