Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Logical Reasoning on CounterBench (test)

88.9Accuracy

FLEx

18.59636.84855.173.352Jan 7, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.01
88.9
2026.01
86.5
2026.01
85.9
2026.01
85.2
2026.01
84.9
2026.01
84.6
2026.01
82.8
2026.01
81.3
2026.01
80.7
2026.01
80.6
2026.01
80.3
2026.01
80.2
2026.01
79.7
2026.01
78.3
2026.01
78.2
2026.01
78
2026.01
76
2026.01
75.9
2026.01
75.8
2026.01
74.7
2026.01
74.6
2026.01
74.3
2026.01
74
2026.01
72.1
2026.01
70.4
2026.01
69.9
2026.01
69.7
2026.01
69
2026.01
68.8
2026.01
68.7
2026.01
68.5
2026.01
68.5
2026.01
66.7
2026.01
65.4
2026.01
64.9
2026.01
61.3
2026.01
58.1
2026.01
55.2
2026.01
53.3
2026.01
51.6
2026.01
51.1
2026.01
50.5
2026.01
49.8
2026.01
49.7
2026.01
49.7
2026.01
48.5
2026.01
47.8
2026.01
47.8
2026.01
47
2026.01
43.4
2026.01
42.6
2026.01
39.3
2026.01
38
2026.01
21.8
2026.01
21.3