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

Visual Perception on HallusionBench

71.08Accuracy

VL-Rethinker

50.300855.695461.0966.4846Apr 12, 2026Apr 20, 2026Apr 29, 2026May 7, 2026May 16, 2026May 24, 2026Jun 2, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
71.08
2026.04
68.63
2026.04
67.72
2026.04
65.8
2026.04
65.5
2026.04
65.19
2026.04
64.9
2026.06
63.85
2026.04
63.83
2026.04
63.6
2026.04
62.57
2026.06
62.36
2026.06
61.41
2026.06
60.88
2026.06
60.46
59.73
2026.06
59.52
59.2
2026.06
58.46
2026.04
57.52
2026.04
56.57
2026.04
56.36
2026.04
56.15
2026.04
51.1