Automatic Speech Recognition on term-rich set 5.01 h benchmark leaderboard · SOTA2 Research

Automatic Speech Recognition on term-rich set 5.01 h

9.19CERBreeze-ASR-25

Evaluation Results

Method	Links
Breeze-ASR-252026.07		9.19	59.51	7.01	82.19	69.03	41.74
Breeze-ASR-252026.07		9.79	60.13	7.63	81.2	69.09	41.38
Breeze-ASR-252026.07		11.37	52.5	—	80.55	63.57	48.66
whisper-l-v3-turbo2026.07		12.82	60.86	6.95	63.79	62.29	45.93
whisper-l-v3-turbo2026.07		13.8	53.9	—	65.11	58.98	49.67
Qwen3-ASR-1.7B2026.07		13.93	56.6	7.57	76.11	64.92	46.26
Qwen3-ASR-1.7B2026.07		14.02	56.56	7.53	74.43	64.27	47.1
whisper-l-v3-turbo2026.07		15.89	59.74	5.84	65.03	62.28	48.77
Qwen3-ASR-1.7B2026.07		16.13	49.03	—	72.84	58.61	53.31
Qwen3-ASR-0.6B2026.07		18.86	53.74	9.45	65.92	59.21	54.64
Qwen3-ASR-0.6B2026.07		18.91	44.28	—	71.19	54.6	58.2
Breeze-ASR-262026.07		21.67	49.73	22.42	63.32	55.71	54.53
Qwen3-ASR-0.6B2026.07		25.39	54.43	10.15	59.79	56.99	64.99
Breeze-ASR-262026.07		26.05	44.89	17.59	57.15	50.28	62.27
Breeze-ASR-262026.07		38.28	27.31	—	56.26	36.77	75.14