Multi-hop QA
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
65.6Exact Match
143
May 25, 2026
77.2EM
95
May 25, 2026
56.4Exact Match (EM)
67
May 13, 2026
57.8Exact Match (EM)
46
May 13, 2026
62EM
42
May 25, 2026
56EM
27
Mar 11, 2026
74.9Accuracy (%)
25
Apr 3, 2026
55.1Pass@4
20
Jun 17, 2026
76.95Cover-EM
20
May 25, 2026
70.5Accuracy
17
Mar 20, 2026
6.1Exact Match (EM)
16
May 11, 2026
18.5Exact Match (EM)
16
May 11, 2026
45.2Exact Match (EM)
14
Jun 16, 2026
50.1EM
14
Jun 16, 2026
87.9F1 Score
14
Mar 4, 2026
59.7F1 Score
11
Apr 21, 2026
37.5Pass@4
10
Jun 17, 2026
64.3Pass@4
10
Jun 17, 2026
36.1Pass@4
10
Jun 17, 2026
59.1Pass@4
10
Jun 17, 2026
57.5EM
10
May 20, 2026
70.4Accuracy
9
Feb 26, 2026
91.5Accuracy
9
Feb 26, 2026
48.2F1 Score
8
Apr 21, 2026
38.2F1
8
Apr 21, 2026