What hardware do I need to run Whisper large-v3?

At its most compressed (Int8) it needs roughly 1.5GB of VRAM, and about 3GB for good quality. VRAM figures are approximate and depend on context length and settings.

Is Whisper large-v3 free for commercial use?

Yes. Whisper large-v3 is licensed under MIT, which permits commercial use with no meaningful conditions.

Can I run Whisper large-v3 on Apple Silicon?

Yes. Whisper large-v3 has builds optimised for Apple Silicon, through MLX or GGUF on a Mac.

Does Whisper large-v3 run on CPU?

Yes, Whisper large-v3 can run on the CPU, though generation is slower than on a GPU.

Whisper large-v3

Name: Whisper large-v3
Author: OpenAI

OpenAI · Speech to text · 1.55B · Released 6 November 2023

Commercial use permitted AudioText Open weights Runs on CPU Apple Silicon

Whisper large-v3 is the standard against which local speech-to-text is measured. Trained on a huge multilingual corpus, it transcribes and translates robustly across accents, noise, and dozens of languages. Locally it is rarely run through OpenAI's own code: the community runtimes whisper.cpp and faster-whisper are far more efficient, running it on a CPU or a modest GPU, with 8-bit options that cut memory further. Its quirks are a tendency to hallucinate during long silences and a 30-second processing window that means long audio is chunked. For real-time use, a smaller or distilled Whisper trades a little accuracy for speed.

Strengths

Strong, robust transcription across many languages
Efficient community runtimes (whisper.cpp, faster-whisper) run it on modest hardware
MIT licensed, with no commercial-use conditions

Weaknesses

Can hallucinate text during long silences or noisy passages
Processes audio in 30-second windows, so very long files need chunking
Real-time use needs a smaller or distilled variant on modest hardware

Hardware requirements

Quantisation	Approx. VRAM	Notes
Int8	~1.5GB	Common for faster-whisper, minimal accuracy loss
FP16	~3GB	Full precision

Also runs on CPU (slower). Optimised builds available for Apple Silicon.

What you'd need to run this

Roughly what a machine to run this would need, at up to three levels of quality. Memory is the deciding factor.

Minimum to run it

Int8 · ~1.5GB needed

One 12GB GPU

NVIDIA GeForce RTX 3060 12GB

or a Mac or mini-PC with unified memory, if you prefer no discrete GPU, Mac mini M5 Pro .

At least 32GB of system RAM alongside the card.

around £700–£1,100

What else 12GB runs →

Best quality

FP16 · ~3GB needed

One 12GB GPU

NVIDIA GeForce RTX 3060 12GB

or a Mac or mini-PC with unified memory, if you prefer no discrete GPU, Mac mini M5 Pro .

At least 32GB of system RAM alongside the card.

around £700–£1,100

What else 12GB runs →

Licence

MIT — read the licence

Benchmarks

Benchmark	Score	Source	As of
Open ASR Leaderboard average WER	7.44	Open ASR Leaderboard paper (independent)	October 2025

Whisper large-v3: common questions

What hardware do I need to run Whisper large-v3?: At its most compressed (Int8) it needs roughly 1.5GB of VRAM, and about 3GB for good quality. VRAM figures are approximate and depend on context length and settings.
Is Whisper large-v3 free for commercial use?: Yes. Whisper large-v3 is licensed under MIT, which permits commercial use with no meaningful conditions.
Can I run Whisper large-v3 on Apple Silicon?: Yes. Whisper large-v3 has builds optimised for Apple Silicon, through MLX or GGUF on a Mac.
Does Whisper large-v3 run on CPU?: Yes, Whisper large-v3 can run on the CPU, though generation is slower than on a GPU.

Availability

Recommended for

Local transcription and subtitling
Multilingual speech-to-text
Feeding audio into a local RAG or notes pipeline

Run it with

Related guides

Speech to text and text to speech

Glossary

Catalogue entry last verified 30 July 2026. Specifications change; verify anything you are about to spend money on.