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.
| # | Model | Params | Context | Commercial use | 30d | Min VRAM |
|---|---|---|---|---|---|---|
| 01 | speaker-diarization-3.1 | — | — | ✓ mit | 8.9 M | — |
| 02 | whisper-large-v3-turbo | 810 M | — | ✓ mit | 8.7 M | from 2.4 GB |
| 03 | whisperkit-coreml | — | — | unknown | 8.3 M | — |
| 04 | whisper-large-v3 | 1.5 B | — | ✓ apache-2.0 | 5.5 M | from 10.9 GB |
| 05 | speaker-diarization-community-1 | — | — | ✓ cc-by-4.0 | 5.3 M | — |
| 06 | whisper-base | 70 M | — | ✓ apache-2.0 | 4.8 M | from 0.8 GB |
| 07 | wav2vec2-large-xlsr-53-portuguese | — | — | ✓ apache-2.0 | 4.6 M | — |
| 08 | Qwen3-ASR-0.6B | 940 M | — | ✓ apache-2.0 | 4.3 M | from 2.7 GB |
| 09 | wav2vec2-large-xlsr-53-russian | — | — | ✓ apache-2.0 | 4.0 M | — |
| 10 | voice-activity-detection | — | — | ✓ mit | 2.7 M | — |
| 11 | wav2vec2-large-xlsr-53-polish | — | — | ✓ apache-2.0 | 2.7 M | — |
| 12 | wav2vec2-large-xlsr-53-japanese | — | — | ✓ apache-2.0 | 2.7 M | — |
| 13 | mms-300m-1130-forced-aligner | 320 M | — | ✗ cc-by-nc-4.0 | 2.5 M | from 1.9 GB |
| 14 | whisper-small | 240 M | — | ✓ apache-2.0 | 2.3 M | from 1.6 GB |
| 15 | Voxtral-Mini-4B-Realtime-2602 | 4.4 B | — | ✓ apache-2.0 | 2.3 M | from 20.7 GB |
| 16 | faster-whisper-small | — | — | ✓ mit | 2.1 M | — |
| 17 | nemotron-3.5-asr-streaming-0.6b-gguf | — | — | other | 2.0 M | from 1.0 GB |
| 18 | Qwen3-ASR-1.7B | 2.4 B | — | ✓ apache-2.0 | 2.0 M | from 6.0 GB |
| 19 | whisper-tiny | 40 M | — | ✓ apache-2.0 | 2.0 M | from 0.7 GB |
| 20 | wav2vec2-large-xlsr-53-dutch | — | — | ✓ apache-2.0 | 2.0 M | — |
| 21 | parakeet-unified-en-0.6b-gguf | — | — | ✓ cc-by-4.0 | 1.9 M | from 1.0 GB |
| 22 | parakeet-tdt-0.6b-v2 | 620 M | — | ✓ cc-by-4.0 | 1.9 M | from 3.3 GB |
| 23 | parakeet-ctc-1.1b | 1.1 B | — | ✓ cc-by-4.0 | 1.9 M | from 2.0 GB |
| 24 | wav2vec2-indonesian-javanese-sundanese | — | — | ✓ apache-2.0 | 1.8 M | — |
| 25 | wav2vec2-base-960h | 90 M | — | ✓ apache-2.0 | 1.7 M | from 0.9 GB |
| 26 | wav2vec2-large-xlsr-53-greek | — | — | ✓ apache-2.0 | 1.7 M | — |
| 27 | distil-large-v3 | 760 M | — | ✓ mit | 1.7 M | from 5.6 GB |
| 28 | faster-whisper-base | — | — | ✓ mit | 1.5 M | — |
| 29 | parakeet-tdt-0.6b-v3 | 630 M | — | ✓ cc-by-4.0 | 1.4 M | from 3.4 GB |
| 30 | wav2vec2-xls-r-300m-cs-250 | 320 M | — | ✓ apache-2.0 | 1.4 M | from 1.9 GB |
| 31 | wav2vec2-large-xlsr-53-arabic | — | — | ✓ apache-2.0 | 1.4 M | — |
| 32 | faster-whisper-tiny | — | — | ✓ mit | 1.3 M | — |
| 33 | wav2vec2-large-xlsr-53-hungarian | — | — | ✓ apache-2.0 | 1.3 M | — |
| 34 | wav2vec2-large-xlsr-53-chinese-zh-cn | — | — | ✓ apache-2.0 | 1.3 M | — |
| 35 | romanian-wav2vec2 | 320 M | — | ✓ apache-2.0 | 1.2 M | from 1.9 GB |
| 36 | faster-whisper-tiny.en | — | — | ✓ mit | 1.2 M | — |
| 37 | faster-whisper-large-v3 | — | — | ✓ mit | 1.2 M | — |
| 38 | Wav2Vec2-large-xlsr-hindi | 320 M | — | unknown | 1.1 M | from 1.9 GB |
| 39 | wav2vec2-large-xlsr-53-th | — | — | cc-by-sa-4.0 | 1.1 M | — |
| 40 | wav2vec2-large-voxrex-swedish | 320 M | — | ✓ cc0-1.0 | 1.1 M | from 1.9 GB |
| 41 | nemotron-3.5-asr-streaming-0.6b | 640 M | — | other | 1.0 M | from 1.4 GB |
| 42 | wav2vec2-large-xlsr-53-telugu | — | — | ✓ apache-2.0 | 1.0 M | — |
| 43 | cohere-transcribe-03-2026 | 2.1 B | — | ✓ apache-2.0 | 1.0 M | from 5.4 GB |
| 44 | cohere-transcribe-03-2026-gguf | — | — | ✓ apache-2.0 | 1,000 K | from 2.2 GB |
| 45 | filipino-wav2vec2-l-xls-r-300m-official | — | — | ✓ apache-2.0 | 982 K | — |
| 46 | wav2vec2-large-xlsr-53-persian | — | — | ✓ apache-2.0 | 972 K | — |
| 47 | wav2vec2-large-xlsr-korean | 320 M | — | ✓ apache-2.0 | 877 K | from 1.9 GB |
| 48 | wav2vec2-large-xls-r-300m-Urdu | 320 M | — | ✓ apache-2.0 | 867 K | from 1.9 GB |
| 49 | w2v-xls-r-uk | 320 M | — | ✓ apache-2.0 | 764 K | from 1.9 GB |
| 50 | wav2vec2-xls-r-300m-hebrew | 320 M | — | unknown | 757 K | from 1.9 GB |
VRAM figures are estimates for the smallest available quantization at 8K context — see /methodology. Downloads refresh nightly from the Hugging Face API.