1,011 models · refreshed nightly

Automatic speech recognition models

Every model in the catalog with its licence, estimated VRAM and daily-tracked downloads. Filters update the URL — share any view.

All text-generation 400 text-to-image 150 ASR 89 image-text-to-text 82 sentence-similarity 55 feature-extraction 47 translation 44 text-classification 39 image-classification 32 token-classification 22 text-to-speech 16 audio-classification 12 image-segmentation 8 object-detection 8 summarization 7

Any RTX 3060 · 12 GB RTX 4070 · 16 GB RTX 4090 · 24 GB 2 × 24 GB Mac M3 · 24 GB Mac M4 Max · 64 GB

89 models

#	Model	Params	Context	Commercial use	30d	Min VRAM
51	vakyansh-wav2vec2-tamil-tam-250 Harveenchadha · ASR	—	—	✓ mit	752 K	—
52	wav2vec2-xls-r-300m-mixed mesolitica · ASR	—	—	unknown	720 K	—
53	parakeet-tdt-0.6b-v2 nvidia · ASR	—	—	✓ cc-by-4.0	719 K	—
54	VibeVoice-ASR microsoft · ASR	8.7 B	—	✓ mit	695 K	from 20.9 GB
55	nb-wav2vec2-1b-nynorsk NbAiLab · ASR	960 M	—	✓ apache-2.0	693 K	from 4.9 GB
56	whisper-tiny mlx-community · ASR	—	—	unknown	679 K	—
57	wav2vec2-large-xlsr-53-finnish jonatasgrosman · ASR	—	—	✓ apache-2.0	665 K	—
58	wav2vec2-base-vi-vlsp2020 nguyenvulebinh · ASR	—	—	✗ cc-by-nc-4.0	665 K	—
59	wav2vec2-xls-r-300m-bengali arijitx · ASR	—	—	✓ apache-2.0	622 K	—
60	whisper-tiny.en Xenova · ASR	—	—	✓ apache-2.0	614 K	—
61	parakeet-tdt-0.6b-v3-gguf handy-computer · ASR	—	—	✓ cc-by-4.0	606 K	from 1.0 GB
62	nb-wav2vec2-1b-bokmaal-v2 NbAiLab · ASR	960 M	—	✓ apache-2.0	581 K	from 4.9 GB
63	Phi-4-multimodal-instruct microsoft · ASR	5.6 B	131 K	✓ mit	572 K	from 15.4 GB
64	wav2vec2-xls-r-parlaspeech-hr classla · ASR	320 M	—	unknown	564 K	from 1.9 GB
65	whisper-bemba-stt AbelZimba · ASR	240 M	—	unknown	546 K	from 1.6 GB
66	wav2vec2-xlsr-nepali gagan3012 · ASR	—	—	✓ apache-2.0	542 K	—
67	granite-speech-3.3-2b ibm-granite · ASR	3.0 B	—	✓ apache-2.0	529 K	from 13.9 GB
68	whisper-medium-gguf handy-computer · ASR	—	—	✓ apache-2.0	518 K	from 1.1 GB
69	faster-whisper-medium Systran · ASR	—	—	✓ mit	514 K	—
70	wav2vec2-xls-r-300m-ftspeech saattrupdan · ASR	320 M	—	other	512 K	from 1.9 GB
71	wav2vec2-xls-r-300m-cv7-turkish mpoyraz · ASR	—	—	✓ cc-by-4.0	490 K	—
72	wav2vec2-large-xlsr-mvc-swahili eddiegulay · ASR	320 M	—	✓ apache-2.0	476 K	from 1.9 GB
73	wav2vec2-xlsr-53-espeak-cv-ft facebook · ASR	—	—	✓ apache-2.0	461 K	—
74	wav2vec2-large-xlsr-catala softcatala · ASR	—	—	✓ apache-2.0	455 K	—
75	vakyansh-wav2vec2-sanskrit-sam-60 Harveenchadha · ASR	—	—	unknown	450 K	—
76	Voxtral-Mini-4B-Realtime-2602-gguf handy-computer · ASR	—	—	✓ apache-2.0	430 K	from 3.6 GB
77	wav2vec2-large-xlsr-53-estonian anton-l · ASR	—	—	✓ apache-2.0	426 K	—
78	granite-speech-4.1-2b ibm-granite · ASR	2.3 B	—	✓ apache-2.0	415 K	from 6.2 GB
79	reverb-diarization-v1 Revai · ASR · gated	—	—	other	413 K	—
80	wav2vec2-large-960h-lv60-self facebook · ASR	—	—	✓ apache-2.0	412 K	—
81	Qwen3-ForcedAligner-0.6B Qwen · ASR	920 M	—	✓ apache-2.0	407 K	from 2.7 GB
82	parakeet-tdt-0.6b-v3-coreml FluidInference · ASR	—	—	✓ cc-by-4.0	400 K	—
83	wav2vec2-large-xlsr-lithuanian m3hrdadfi · ASR	—	—	✓ apache-2.0	395 K	—
84	wav2vec2-large-xlsr-kazakh aismlv · ASR	320 M	—	✓ apache-2.0	386 K	from 1.9 GB
85	wav2vec2-xls-r-300m-sk-cv8 comodoro · ASR	—	—	✓ apache-2.0	383 K	—
86	speaker-diarization pyannote · ASR · gated	—	—	✓ mit	373 K	—
87	GigaAM-v3 ai-sage · ASR	—	—	✓ mit	370 K	—
88	whisper-medium openai · ASR	760 M	—	✓ apache-2.0	364 K	from 4.0 GB
89	wav2vec2-large-xlsr-malayalam gvs · ASR	—	—	✓ apache-2.0	350 K	—

VRAM figures are estimates for the smallest available quantization at 8K context — see /methodology. Downloads refresh nightly from the Hugging Face API.