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

Visual Perception and Reasoning on V*

90.1Overall Accuracy

DeepEyes‡

64.15270.888577.62584.3615Mar 29, 2026Apr 5, 2026Apr 12, 2026Apr 20, 2026Apr 27, 2026May 4, 2026May 12, 2026
Updated 21d ago

Evaluation Results

MethodLinks
2026.03
90.191.388.2
2026.03
89.591.386.8
2026.03
88.590.485.5
2026.03
88.2--
2026.03
8887.888.2
2026.03
85.984.388.2
2026.03
84.382.686.8
2026.04
83.884.382.9
2026.05
83.7584.2683.24
2026.05
83.384.481.6
2026.05
82.783.581.6
2026.04
82.282.681.6
2026.05
81.981.9481.86
2026.05
81.7681.2482.28
2026.03
80.682.677.6
2026.05
80.683.2677.94
2026.05
80.683.576.3
2026.05
80.681.779.8
2026.03
80.178.382.9
2026.04
80.182.676.3
2026.05
80.179.181.6
2026.04
79.681.776.3
2026.04
79.681.776.3
2026.05
79.181.0577.15
2026.05
79.181.775
2026.05
78.2578.3978.11
2026.05
7879.176.3
2026.05
77.478.376.3
2026.05
76.6577.1274.35
2026.03
75.975.776.3
2026.05
75.48068.5
2026.03
73.973.175
2026.03
72.87372.4
2026.04
67.572.260.5
2026.05
67.572.260.5
2026.05
65.1569.6858.39