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

Multimodal Benchmarking on MMBench EN

80.5Accuracy (MMB)

Vanilla

62.92467.48772.0576.613Apr 20, 2026Apr 29, 2026May 9, 2026May 19, 2026May 28, 2026Jun 7, 2026Jun 17, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.04
80.510010031.976.6-
2026.04
80.4106.694.818.829.8-
2026.04
8082.174.319.532.6-
2026.04
79.981.374.621.939-
2026.04
79.782.478.620.634.5-
2026.04
79.292.480.716.420.6-
2026.04
78.174.26818.827.6-
2026.04
77.87574.71825-
2026.04
77.570.674.813.135.6-
2026.04
77.476.665.31723-
2026.04
73.967.554.112.329.6-
2026.04
73.961.662.512.620.2-
2026.04
72.665.954.21230.4-
2026.04
70.667.15811.729.7-
2026.06
69.1----98.3
2026.06
68.6----100.7
2026.06
68.6----99.1
2026.06
68.5----100
2026.04
67.757.3538.419.1-
2026.06
67.3----95.6
2026.06
67----97.2
2026.04
66.860.953.28.219.6-
2026.06
66.6----96.3
2026.06
66.5----97.1
2026.06
66.2----95.2
2026.04
63.95353.87.619-
2026.06
63.6----89.4