General Audio Understanding on MMSU 1.0 (test) benchmark leaderboard · SOTA2 Research

General Audio Understanding on MMSU 1.0 (test)

72.13Perception SemanticsQwen3-Omni

Evaluation Results

Method	Links
Qwen3-Omni2026.01		72.13	55.83	38.85	53.2	86.64	82.19	43.58	78.88	65.63
Qwen2.5-Omni2026.01		61.11	43.96	33.2	43.97	82.4	76.77	46.87	75.21	59.09
DIFFA-22026.01		60.63	39.04	41.92	45.58	85.29	77.58	43.58	76.4	60.45
GPT-4o-Audio2026.01		59.7	41.56	21.44	39.67	80.83	78.74	26.25	71.96	56.38
DIFFA-22026.01		59.69	39.57	40.93	45.06	85.02	77.28	42.09	76	60.1
Kimi-Audio2026.01		57.64	42.3	35.74	43.52	81.77	76.65	55.22	76.03	59.28
Qwen-Audio-Chat2026.01		57.21	38.52	24.7	35.69	58.61	59.78	25.6	55.93	46.92
MiniCPM-O2026.01		56.56	34.05	36.48	40.54	80.71	74.72	46.71	73.57	56.53
DIFFA2026.01		52.67	36.65	35.12	40.28	81.53	72.68	45.67	72.92	56.04
Qwen2-Audio2026.01		52.14	32.87	35.56	39.02	77.62	64.81	46.67	68.9	53.27
Gemini 2.0 Flash2026.01		47.17	41.3	30.62	40.83	70.69	70.69	36.16	47.83	51.03
Baichuan-Audio2026.01		39.63	31.26	27.09	31.48	57.96	63.92	34.35	55.7	43.09
Phi-4-multimodal2026.01		38.72	34.86	29.56	33.41	57.81	65.94	42.09	57.59	44.96
Salmonn2026.01		31.55	29.08	28.71	29.83	36.43	26.22	25.26	30.04	30.01
GLM-4-Voice2026.01		27.8	24.52	27.34	26.18	46.1	48.16	44.35	46.76	35.51
LTU2026.01		21.34	22.46	18.73	20.81	22.65	25.53	24.74	24.37	22.61