Spoken Dialogue on URO-Bench Chinese Basic Track benchmark leaderboard · SOTA2 Research

Spoken Dialogue on URO-Bench Chinese Basic Track

99.83Repeat ScoreStep-Audio 2 mini

Evaluation Results

Method	Links
Step-Audio 2 mini2026.02		99.83	81.18	80.67	64.71	70.02	45.09	74.94	74.29	72.58	79.58
Qwen3-Omni2026.02		99.48	84.54	88.67	52.07	73.02	26.67	36.09	79.59	78.86	74.54
Covo-Audio-Chat-FD2026.02		98.35	80.72	83.67	75.16	79.37	70.7	78.85	84.9	83.93	88.34
Covo-Audio-Chat2026.02		98.17	79.89	88	77.34	83.6	68.42	80.69	90.02	87.18	90.41
Covo-Audio-Chat2026.02		97.68	79.22	91	76.91	80.6	64.21	79.31	89.34	87.99	89.28
Fun-Audio-Chat2026.02		97.5	80.95	79	51.85	67.9	44.91	76.09	80.59	86.46	84.62
Step-Audio 22026.02		96.16	79.16	—	—	74.06	—	62.99	76.23	84.54	—
GPT-4o Audio2026.02		93.5	81.6	88	42.67	76	25.33	81.33	86.4	82.93	80
GLM-4-Voice2026.02		92.64	77.08	69	28.75	56.96	15.78	78.85	83.35	82.12	84.48
Kimi-Audio2026.02		73.32	78.4	—	—	66.93	—	52.45	69.9	81.81	—
Step-Audio-AQAA2026.02		63.72	68.33	—	—	47.21	—	67.4	56.23	31.5	—
Qwen2.5-Omni2026.02		19.66	73.22	—	—	67.7	—	61.76	70.4	79.71	—
Freeze-Omni2026.02		4.97	71.82	7.66	9.58	16.4	11.75	47.35	67.98	64.89	71.28
Freeze-Omni2026.02		3.71	42.75	2.33	4.58	4.94	12.81	47.13	66.17	59.94	70.43
MiMo-Audio2026.02		0.44	79.61	92.67	35.73	83.6	61.05	38.16	85.9	89.1	84.84