Multi-hop Reasoning
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
41.62F1 Score
68
May 14, 2026
95.6Accuracy
50
May 29, 2026
51Accuracy
48
Jun 2, 2026
53EM
41
Feb 26, 2026
75.05Accuracy / F1
30
Jun 4, 2026
52.8ACC
27
Feb 26, 2026
54.6ACC
27
Feb 26, 2026
25.59ACC
27
Feb 26, 2026
21.14Accuracy/F1
26
Jun 4, 2026
0.7AUROC
26
Feb 26, 2026
67.19AUROC
26
Feb 26, 2026
59.64Accuracy
24
Jun 17, 2026
72.2Accuracy
23
Jun 2, 2026
55F1 Score
22
Jun 2, 2026
81F1 Score
21
Jun 2, 2026
70.31Exact Match (EM)
17
Mar 24, 2026
95.4Accuracy (32K Context)
17
Mar 4, 2026
82.67Accuracy
15
Apr 29, 2026
71.84F1 Score
15
Apr 29, 2026
30.08F1 Score
15
Apr 29, 2026
73.3F1 Score
15
Apr 29, 2026
80.8Exact Match
15
Feb 27, 2026
70Exact Match
15
Feb 27, 2026
1Relative Cost
14
Apr 8, 2026
1Relative Cost
14
Apr 8, 2026