Question Answering on HotpotQA (test) (EM and F1 Metrics) benchmark leaderboard · SOTA2 Research

Question Answering on HotpotQA (test) (EM and F1 Metrics)

56.2EMVEROIC

Evaluation Results

Method	Links
VEROIC2026.04		56.2	72.8
RISK PREDICTOR2026.04		55.9	72.3
VEROIC2026.04		55.8	72.1
CONFIDENCE-THRESHOLD2026.04		55.6	71.9
OUTPUT-TRIGGER2026.04		55.3	71.6
RISK PREDICTOR2026.04		55.3	71.9
CONFIDENCE-THRESHOLD2026.04		54.9	71.6
PERIODIC2026.04		54.7	70.8
OUTPUT-TRIGGER2026.04		54.6	71.3
RANDOM2026.04		53.8	70.6
PERIODIC2026.04		53.3	70.3
RANDOM2026.04		53.1	70.2
FIXED2026.04		51.6	68.7
FIXED2026.04		51.2	68.3
R3AG2026.04		29.13	40.73
IRCoT2026.04		29.05	40.52
Adaptive-RAG2026.04		28.87	40.54
BGE-large2026.04		27.54	38.6
Oracle Single Best2026.04		27.54	38.6
E5-large2026.04		26.87	37.87
Qwen3-embedding-4b2026.04		26.64	37.62
E5-base-v22026.04		25.31	35.75
BGE-m32026.04		24.73	35.01
Qwen3-embedding-0.6b2026.04		24.42	34.63
RouterRetriever2026.04		24.37	34.98
AAR-contriever-kilt2026.04		23.04	33.37
DIVER-Retriever-0.6b2026.04		22.75	32.78
Random2026.04		22.13	32.09
FLARE2026.04		20.77	28.03
LTRR2026.04		19.86	28.64
Naive2026.04		19.16	26.96
BM252026.04		17.87	26.15