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

Multi-modal Task Evaluation on MM-MergeBench Seen Tasks

83.74SciQA Score

Individual

60.849666.792372.73578.6777Mar 6, 2026
Updated 4mo ago

Evaluation Results

MethodLinks
2026.03
83.7496.0267.5843.465.564.857.2975.5469.23
2026.03
77.5477.0163.545.7240.7542.3454.7453.2556.86
2026.03
77.0573.3365.8747.8859.6645.2355.0652.9259.63
2026.03
76.974.0867.0535.9865.3766.6756.0966.8763.62
2026.03
75.6678.6159.4841.9241.7141.1952.2345.9754.6
2026.03
73.4365.5467.244.862.9746.6152.845.957.33
2026.03
71.9457.4967.0638.962.8744.849.239.2153.93
2026.03
71.5957.2566.2639.3862.5644.9349.1339.5953.84
2026.03
71.4955.8866.7339.6765.1244.3547.0634.4653.09
2026.03
71.157.8267.5938.2264.3544.5848.937.0153.7
2026.03
69.4678.456.8438.9538.5840.2351.1338.8251.55
2026.03
61.7340.8762.8836.141.1641.0349.0714.0943.37