Misleading Omission Detection on MM-MISLEADING benchmark leaderboard · SOTA2 Research

Misleading Omission Detection on MM-MISLEADING

86AccuracyQwen3-VL-235B

Evaluation Results

Method	Links
Qwen3-VL-235B2026.01		86	80	96	88	96	76	85
OMGUARD2026.01		86	87	84	85	84	88	86
OMGUARD2026.01		84	79	93	85	91	75	82
LLama3-VL-90B2026.01		72	67	85	75	80	58	67
Qwen3-VL-8B2026.01		70	70	68	69	69	71	70
GLM-4.1V-9B2026.01		70	74	60	66	67	79	72
Qwen3-VL-8B2026.01		68	62	95	75	89	42	57
InternVL3.5-8B2026.01		64	59	95	73	86	34	49
LLama3-VL-11B2026.01		51	51	58	54	51	43	47
Llava-7B2026.01		50	50	100	67	0	0	0