Watermarking Robustness on LVLM Evaluation Set (test) benchmark leaderboard · SOTA2 Research

Watermarking Robustness on LVLM Evaluation Set (test)

0Relative DropKGW

Evaluation Results

Method	Links
KGW2026.02		0	100
SynthID2026.02		0	99.9
IE2026.02		0	99.66
MorphMark2026.02		0	99.69
VLA-Mark2026.02		0	99.61
AGMark2026.02		0	99.89
KGW2026.02		0.07	99.93
KGW2026.02		0.11	99.89
KGW2026.02		0.2	99.8
KGW2026.02		0.24	99.76
AGMark2026.02		0.7	99.29
AGMark2026.02		0.71	99.28
AGMark2026.02		0.82	99.07
MorphMark2026.02		0.83	98.86
AGMark2026.02		0.87	99.02
MorphMark2026.02		1.52	98.17
VLA-Mark2026.02		1.71	97.91
IE2026.02		1.76	97.91
VLA-Mark2026.02		1.82	97.8
IE2026.02		1.86	97.81
IE2026.02		1.96	97.71
MorphMark2026.02		2.08	97.62
MorphMark2026.02		2.95	96.75
VLA-Mark2026.02		3.34	96.28
IE2026.02		3.55	96.12
VLA-Mark2026.02		4.13	95.5
SynthID2026.02		6.66	93.24
SynthID2026.02		6.97	92.93
SynthID2026.02		9.06	90.84
SynthID2026.02		10.46	89.44
AGMark2026.02		11.28	88.61
KGW2026.02		11.84	88.16
IE2026.02		13.25	86.46
MorphMark2026.02		16.46	83.28
VLA-Mark2026.02		26.36	73.35
SynthID2026.02		36.32	63.58