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

Multimodal Reasoning on MMBench (Accuracy and Efficiency)

64.35MMBench Accuracy

PDrop

59.077260.446161.81563.1839May 10, 2026
Updated 22d ago

Evaluation Results

MethodLinks
2026.05
64.3599.284.72105.96
2026.05
64.1899.634.68104.45
2026.05
64.091008.54167.5
2026.05
64.0999.594.68136.4
2026.05
64.0998.624.51125.75
2026.05
64.0899.135.02109.12
2026.05
63.9299.793.4489.93
2026.05
63.7598.393.8692.07
2026.05
63.7599.254.7125.8
2026.05
63.6697.812.968.73
2026.05
63.6697.773.2594.62
2026.05
63.5798.642.677.94
2026.05
62.9798.323.4484.22
2026.05
62.6397.083.4681.94
2026.05
62.4696.992.4280.47
2026.05
62.3798.533.42113.78
2026.05
61.6896.82.6274.28
2026.05
61.695.772.6867.59
2026.05
59.2896.22.5992.8