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

Real-world Visual Understanding on RealWorldQA (Score Metric)

72.29Score

Ours-Mixed

-2.2405617.1087236.45855.80728Oct 27, 2025Nov 30, 2025Jan 3, 2026Feb 6, 2026Mar 12, 2026Apr 15, 2026May 20, 2026
Updated 13d ago

Evaluation Results

MethodLinks
2026.03
72.29
2026.03
71.63
2026.03
71.11
2026.03
70.46
2026.03
69.54
2026.03
68.76
2026.05
68.6
2026.03
68.5
2026.05
68.5
2026.05
68.4
2026.05
68
2026.03
67.97
2026.05
67.4
2026.05
67.3
2026.05
66.8
2026.05
66.5
2026.03
64.97
2026.05
64.7
2026.05
63.9
2026.03
63.01
2026.03
62.88
2026.03
62.88
2026.03
61.7
2026.03
60.92
2025.10
59.78
2025.10
58.86
2026.03
58.43
2026.03
57.91
2025.10
57.52
2025.10
57.25
2025.10
57.25
2025.10
57.12
2026.03
56.73
2025.10
56.6
2026.03
46.27
2026.03
0.664
2026.03
0.639
2026.03
0.634
2026.03
0.626