whatmodel.sh / stt / Distil-Whisper distil-large-v3.5

Distil-Whisper distil-large-v3.5

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Distil-Whisper distil-large-v3.5 is an open-weights speech-to-text model from Hugging Face (Distil-Whisper), released March 2025. Its published size is 0.76B parameters. Licence: mit.

57.5±1.4#51 of 162 speech-to-text modelsfrom 17 benchmark results across 4 sources

Score within the Open ASR English short-form WER (March 2026) comparison set only.

$ how it compares

The speech-to-text models ranked either side of Distil-Whisper distil-large-v3.5, out of 176 with a score, then more from Hugging Face (Distil-Whisper).

#modelscorewhereRAMreleased
58Omnilingual ASR LLM 7B Meta58±2local~19GB2025-11
59Omnilingual ASR LLM 7B v2 Meta58±3local~19GB2025-12
60Distil-Whisper distil-large-v3.5 Hugging Face (Distil-Whisper)58±1local~2.3GB2025-03
61Whisper large-v3-turbo OpenAI57±1api $0.0005/audio min +local~2.4GB2024-10
62Omnilingual ASR LLM 3B Meta57±2local~11GB2025-11
more from Hugging Face (Distil-Whisper)
77Distil-Whisper distil-large-v3 Hugging Face (Distil-Whisper)56±2local~2.3GB2024-03
122Distil-Whisper distil-large-v2 Hugging Face (Distil-Whisper)48±2local~2.3GB2023-10
123Distil-Whisper distil-small.en Hugging Face (Distil-Whisper)48±4local~0.9GB2023-12

see all speech-to-text models · find the ones that fit your machine

$ what it's good at

long-form transcription▁▁▄▁·▁▂█▄▁▁▁ #15 of 27 * · 7.2% of its score · 1 test

Beats Parakeet TDT 1.1B, Parakeet RNNT 1.1B here, though they rank higher overall. Trails Whisper large (v1), Whisper large-v2, Rev AI Machine (Reverb ASR) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Open ASR English long-form WER9.7% ✎▁▁▅▁·▃▅█▆▁▂▂ beats 13/30avg 12%▇▇▇ 7.2%typical
other tests16 not in a task · 92% of its score
testresultvs the others in its rundifficultycountswhy
AfriSpeech-MultiBench · 7-region African-accent mean WER22.3% ✎▂····▂···▂▆█ beats 10/11avg 46%▇▇▇▇▇ 14%typical
AfriSpeech-MultiBench · Parliament (legal domain) WER18.0% ✎▂▂····▂·▅·▃█ beats 12/12avg 52%▇▇▇▇▇ 12%sharp
Open ASR English short-form WER (March 2026)7.2% ✎▁▁▁▃··▁▁▂██▆ beats 54/86avg 8.2%▇▇▇ 7.3%blunt
SPGISpeech test WER lab2.9% ✎▄▄··▄██··▄██ beats 6/11avg 5.3%▇▇▇ 6.9%sharp
Meanwhile long-form WER lab6.8% ✎█·······██·█ beats 2/3avg 11%▇▇▇ 6.3%sharp
TED-LIUM long-form WER lab4.6% ✎█······█·█·█ beats 1/3avg 6.9%▇▇ 5.9%sharp
RW-Voice-EQ ASR · 4 real-world English sets, mean WER9.7% ✎█▃▃▅▃···▃··▃ beats 0/9avg 15%▇▇ 5.8%typical
AMI WER (Open ASR / ESB test split) lab14.6% ✎▃█▃▃·····▅·▃ beats 5/8avg 24%▇▇ 5.5%typical
8 more tests
testresultvs the others in its rundifficultycountswhy
Earnings-22 WER (short-form, Open ASR / ESB test split) lab11.3% ✎▃······██▃·▃ beats 6/8avg 18%▇▇ 5.3%typical
Rev16 long-form WER lab14.0% ✎█········███ beats 1/3avg 20%▇▇ 4.9%typical
GigaSpeech test WER lab9.8% ✎▁····▁··▁▄█▂ beats 11/17avg 15%▇▇ 4.6%typical
Earnings-21 long-form WER lab10.6% ✎█········███ beats 1/3avg 15%▇▇ 4.6%typical
Earnings-22 long-form WER lab14.2% ✎█······█·█·█ beats 2/3avg 20%▇▇ 4.3%typical
LibriSpeech test-other WER lab5.0% ✎▁▁▁▂▅▃▅▄█▂▂▁ beats 23/54avg 6.6%▇▇ 3.9%typical
LibriSpeech test-clean WER lab2.4% ✎▁·▁·▁▁▁▁▃▇█▆ beats 37/100avg 2.6%▇ 1.2%blunt
TED-LIUM test WER (short-form) lab3.6% ✎▁·▁▇█▃·▁▁▃█▄ beats 23/27avg 5.3%▇ 0.3%blunt

$ where to run it

Other ways to run it

  1. Run it as published · needs ~2.32 GB, NVIDIA only · from github.com/Vaibhavs10
    uvx insanely-fast-whisper --file-name audio.mp3 --model-name distil-whisper/distil-large-v3.5

Official weights and documentation: huggingface.co/distil-whisper ↗

$ how the score was worked out · distil-whisper/distil-large-v… 54%, AfriSpeech-MultiBench (arXiv:… 25%, Open ASR Leaderboard 15% and 1 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
distil-whisper/distil-large-v3.5 model card · lab-reported12▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 54%MIT
AfriSpeech-MultiBench (arXiv:2511.14255v1) · African-accented English ASR benchmark suite2▇▇▇▇▇▇▇▇ 25%CC BY 4.0
Open ASR Leaderboard2▇▇▇▇ 15%Apache-2.0
RW-Voice-EQ Bench (Hume AI)1▇▇ 6%CC BY 4.0
prior · pull toward the middle▇ 1%
weighted average= 57.5

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json