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

Visual Perception and Reasoning on BLINK

70.02Accuracy

CGC-8B

38.383246.596654.8163.0234Feb 24, 2026Mar 11, 2026Mar 27, 2026Apr 12, 2026Apr 27, 2026May 13, 2026May 29, 2026
Updated 2d ago

Evaluation Results

MethodLinks
2026.04
70.02
68.33
2026.02
63
2026.04
57.23
2026.04
57.14
2026.04
56.92
2026.02
56.7
2026.02
56.3
2026.04
56.23
2026.04
56.1
2026.04
56.07
2026.02
56
2026.02
55.9
55.81
2026.04
55.81
2026.02
55.7
2026.04
55.55
55.54
2026.04
54.81
2026.04
54.81
2026.04
54.5
2026.04
54.49
2026.04
54.39
2026.04
53.6
2026.04
53.6
2026.04
53.55
2026.02
52.5
52.4
2026.04
51.53
2026.04
51.36
2026.04
51.1
2026.04
51.08
2026.02
51
2026.04
50.98
2026.04
50.71
2026.04
50.23
2026.04
49.91
2026.04
49.72
2026.04
49.38
2026.04
49.34
2026.04
49.13
2026.05
49.01
2026.05
48.92
2026.04
48.75
2026.04
48.56
2026.05
48.31
2026.05
47.99
2026.05
47.89
2026.05
47.75
2026.05
47.41
2026.05
47.25
2026.05
47.2
2026.04
47.19
2026.05
47.09
2026.04
46.94
2026.05
46.83
2026.05
46.63
2026.04
46.6
2026.04
46.44
2026.04
46.25
2026.04
45.01
2026.05
44.82
2026.04
42.81
39.6