Distil-Whisper large-v3
Distil-Whisper large-v3 is a speech-to-text model (large-v3 distilled) from Hugging Face / Distil-Whisper, released under the MIT license. The reviewed source lists English, and the model is designed for batch transcription, with live use only through a runtime or chunking wrapper. Among the common setups we test, it ranks highest for subtitles with timing (#3 of 26 in the matcher).
English-only distilled Whisper model aimed at faster inference while retaining strong long-form transcription quality.
Key facts
| Provider | Hugging Face / Distil-Whisper |
|---|---|
| Model family | Distil-Whisper |
| Parameters | large-v3 distilled |
| License | MIT · Permissive — license text ↗ |
| Language coverage | English |
| Streaming | Streaming via a runtime or chunking wrapper |
| Timestamps | Yes — in the model or its runtime family |
| Speaker diarization | No — needs a separate component |
| Speech translation | No |
| Documented runtimes | Hugging Face Transformers, openai-whisper, whisper.cpp, faster-whisper, Transformers.js — how to run it |
| Primary source | huggingface.co/distil-whisper/distil-large-v3 ↗ |
| Last reviewed |
Hardware and workload fit
These are Speech Model Match's conservative 1–5 editorial ratings, relative to the other 19 reviewed models. They come from model size, official memory guidance and documented runtimes — not from a universal benchmark. How the ratings work.
| Dimension | Rating | Scale |
|---|---|---|
| CPU-only fit | 2/5 | |
| Apple Silicon fit | 4/5 | |
| NVIDIA GPU fit | 5/5 | |
| Speed | 5/5 | |
| Accuracy (relative) | 5/5 | |
| Low memory use | 3/5 | |
| Long recordings | 5/5 | |
| Live / low latency | 3/5 | |
| Subtitle work | 5/5 | |
| Edge / low-power devices | 2/5 |
Where it is strong
- Apple Silicon fit: rated 4/5
- NVIDIA GPU fit: rated 5/5
- Speed: rated 5/5
- Accuracy (relative): rated 5/5
- Long recordings: rated 5/5
- Subtitle work: rated 5/5
- Timestamps are available in the model or its runtime family
What to check first
- CPU-only fit: rated 2/5
- Edge / low-power devices: rated 2/5
- No native speaker diarization — pair it with a separate diarization pipeline if you need speaker labels
- English-only: not suitable for other languages
How it ranks in common setups
Ranks come from the same deterministic matcher that runs on the homepage, applied to all 26 reviewed models. They are decision guidance, not accuracy claims.
| Setup | Rank | Fit score | Top match |
|---|---|---|---|
| Long recordings on an 8–12 GB NVIDIA GPU, balanced | #4 of 26 | 79/100 · Strong fit | Whisper turbo |
| Live captions or voice input (native streaming + timestamps, speed first) | #8 of 26 | 81/100 · Strong fit | Nemotron 3.5 ASR Streaming 0.6B |
| Subtitles with timing, accuracy first | #3 of 26 | 87/100 · Excellent fit | Whisper medium |
| Meetings with speaker labels (diarization + timestamps) | #6 of 26 | 84/100 · Strong fit | VibeVoice-ASR 7B |
| CPU-only machine, low memory use | #13 of 26 | 61/100 · Possible fit | VibeVoice-ASR-BitNet |
| Apple Silicon with 8–16 GB unified memory, long recordings | #11 of 26 | 69/100 · Possible fit | Whisper small |
| Japanese live transcription (native streaming, speed first) | Not ranked (language not listed) | — | Nemotron 3.5 ASR Streaming 0.6B |
Supported languages
- English
en
Compare coverage across every reviewed model in the language support matrix.
Community activity on Hugging Face
- Downloads, last 30 days
- 740,910
- Likes
- 386
- Repository last updated
- 21 Apr 2026
- Published on the Hub
- 21 Mar 2024
Figures from the Hugging Face Hub API, 11 Nov 2026. Source: huggingface.co/distil-whisper/distil-large-v3 ↗. Downloads count file downloads, not active users, and do not affect the matcher's scores. Compare all models.
How to run Distil-Whisper large-v3 locally
Distil-Whisper keeps Whisper's format, so the main Whisper runtimes can load it. Runtimes below are the ones the official model card or repository documents for this model (checked November 2, 2026).
- Hugging Face Transformers ↗Framework · reference usage on the model card
- openai-whisper ↗Official package · documented on the model card
- whisper.cpp ↗Third-party runtime · documented on the model card
- faster-whisper ↗Third-party runtime · listed in the faster-whisper README
- Transformers.js ↗Framework · JavaScript, documented on the model card
Install command shown in the official huggingface.co page ↗:
pip install --upgrade transformers accelerate datasets[audio]
Audio input: English only. For long recordings the model card recommends chunked inference with 25-second chunks.
Whatever runtime you choose, measure speed and accuracy on your own audio with the benchmark tool; results differ a lot between runtimes, precision and hardware.
Before you deploy Distil-Whisper large-v3
A printable checklist built from the facts on this page. Tick it off before putting the model into production.
- Language. Confirm your language is in the 1-language list on this page (this model is single-language), and test any accents or code-switching you expect.
- License. MIT: Include the copyright notice and the permission notice in all copies or substantial portions. Provided as is, without warranty. License guide.
- Install. Start from the official command:
pip install --upgrade transformers accelerate datasets[audio](full run notes). - Audio input. English only. For long recordings the model card recommends chunked inference with 25-second chunks.
- Hardware. No vendor memory figure is published; our ratings are CPU 2/5, Apple 4/5, NVIDIA 5/5.
- Features you rely on. For Distil-Whisper large-v3, streaming only through a runtime or chunking wrapper; timestamps are documented; speaker labels need a separate diarization component. Test each one you need with real audio.
- Measure. Run 10–20 representative clips, score them with the WER calculator and record speed in the benchmark workspace before you commit.
Alternatives and related models
Similar fit profile: Whisper turbo (OpenAI, 809M), Parakeet TDT 0.6B v3 (NVIDIA, 600M), Qwen3-ASR 0.6B (Qwen, 0.6B). Similarity is computed from the ten fit ratings plus streaming and diarization support.
| Fact | Distil-Whisper large-v3 | Whisper turbo | Parakeet TDT 0.6B v3 |
|---|---|---|---|
| Provider · size | Hugging Face / Distil-Whisper · large-v3 distilled | OpenAI · 809M | NVIDIA · 600M |
| License | MIT Permissive | MIT Permissive | CC BY 4.0 Attribution required |
| Languages | English | 99+ listed languages | 25 European languages |
| Streaming | Via runtime | Via runtime | Via runtime |
| Timestamps | Yes | Yes | Yes |
| Speaker diarization | Separate pipeline | Separate pipeline | Separate pipeline |
| CPU · Apple · NVIDIA fit | 2/5 · 4/5 · 5/5 | 2/5 · 4/5 · 5/5 | 2/5 · 3/5 · 5/5 |
| Live · long recordings | 3/5 · 5/5 | 3/5 · 5/5 | 2/5 · 5/5 |
| Main runtime | Hugging Face Transformers | openai-whisper | NVIDIA NeMo |
Head-to-head pages: Whisper turbo vs Distil-Whisper large-v3
Compare Distil-Whisper large-v3 side by side
Check Distil-Whisper large-v3 for your setup
Pick your language, hardware and workload to see this model's fit score, the reasons behind it and where it ranks against all reviewed models. Everything runs in your browser.
Common questions
Can I use Distil-Whisper large-v3 commercially?
Permissive license; commercial use is allowed under the listed terms. The listed license is MIT; read the current license text ↗ before shipping a product, because terms can change. The license guide compares the obligations of all 26 reviewed models.
Does Distil-Whisper large-v3 support real-time streaming?
Not natively. Live use relies on a runtime or chunking wrapper around a batch model, which adds latency and can split words at chunk borders. If live latency is critical, compare it with a native streaming model such as Nemotron 3.5 ASR Streaming 0.6B.
Which languages does Distil-Whisper large-v3 transcribe?
English only. For other languages pick a multilingual model; the language support matrix shows which reviewed models list each language.
Can Distil-Whisper large-v3 tell speakers apart?
Not natively. Combine it with a dedicated diarization component and evaluate that component separately; the meetings guide covers this workflow.
Primary sources
Reviewed against primary sources on September 16, 2026. Model behavior, runtime support, language coverage and licenses can change; re-check the linked primary source before production use. See the full source directory.