Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Question Answering on ARC (accuracy)

94.2Accuracy

GPT-5 nano

20.900839.930458.9677.9896May 8, 2026
Updated 14d ago

Evaluation Results

MethodLinks
2026.05
94.2
2026.05
93
2026.05
92.1
2026.05
92.1
2026.05
90.3
2026.05
89.9
2026.05
80.2
2026.05
78.1
2026.05
75.3
2026.05
74.1
2026.05
70.3
2026.05
69.8
2026.05
65.6
2026.05
58.19
2026.05
53.5
2026.05
51.45
2026.05
50.34
2026.05
48.81
2026.05
47
2026.05
29.61
2026.05
29.29
2026.05
25.9
2026.05
25.61
2026.05
23.72