For usage instructions follow openai/whisper-large-v3. Large-v3 finetune trained as a baseline with smaller checkpoints in progress. Expecting worse long form and equal short form. Benchmarks. Has occasional repetition issue compared to previous models but achieves competitive/SOTA CER across all tested sets.
Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation, proposed in the paper Robust Speech Recognition via Large-Scale Weak Supervision by Alec Radford et al. from OpenAI.
Parameters1.5B
Context—
Weights24.7 GB
Licenseapache-2.0
AccessOpen weights
Monthly Downloads4.8M
Runs On
What it takes to serve whisper-large-v3 (1.5B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 3.1 GB | 3.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 1.5 GB | 1.9 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 0.8 GB | 0.9 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.
Model Card
By OpenAI, published under apache-2.0, revision 06f233fe06e7.
Whisper
Whisper is a state-of-the-art model for automatic speech recognition (ASR) and speech translation, proposed in the paper Robust Speech Recognition via Large-Scale Weak Supervision by Alec Radford et al. from OpenAI. Trained on >5M hours of labeled data, Whisper demonstrates a strong ability to generalise to many datasets and domains in a zero-shot setting.
Whisper large-v3 has the same architecture as the previous large and large-v2 models, except for the following minor differences:
- The spectrogram input uses 128 Mel frequency bins instead of 80
- A new language token for Cantonese
The Whisper large-v3 model was trained on 1 million hours of weakly labeled audio and 4 million hours of pseudo-labeled audio collected using Whisper large-v2 . The model was trained for 2.0 epochs over this mixture dataset.
Configuration
- Architecture
- WhisperForConditionalGeneration
- Layers
- 32
- Vocabulary size
- 51,866
- Stored precision
- float16
- Model type
- whisper
Identity and Version
- Repository
- openai/whisper-large-v3
- Publisher
- OpenAI
- Task
- Speech recognition
- Modality
- Audio
- Library
- transformers
- Parameters
- 1.5B parameters
- Languages
- en, zh, de, es, ru, ko, fr, ja
- Revision
- 06f233fe06e710322aca913c1bc4249a0d71fce1
- First published
- 2023-11-07
- Last updated
- 2024-08-12
Files and Weights
21 files, 24.7 GB in total. The weights are 7 files totalling 24.7 GB in bin, msgpack, safetensors.
Weights7 files · 24.7 GB
Configuration8 files · 330.8 KB
Tokenizer4 files · 4.3 MB
Documentation1 file · 21.8 KB
Repository1 file · 1.5 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| flax_model.msgpack | Weights | 6.2 GB | e9c7b745947d |
| model.fp32-00001-of-00002.safetensors | Weights | 5.0 GB | 08e0005225b3 |
| model.fp32-00002-of-00002.safetensors | Weights | 1.2 GB | 630ca7746728 |
| model.safetensors | Weights | 3.1 GB | a8e94b85976e |
| pytorch_model.bin | Weights | 3.1 GB | 8f3dd0108a56 |
| pytorch_model.fp32-00001-of-00002.bin | Weights | 5.0 GB | fed9e023b351 |
| pytorch_model.fp32-00002-of-00002.bin | Weights | 1.2 GB | aadd3bdfecf9 |
| added_tokens.json | Configuration | 34.6 KB | — |
| config.json | Configuration | 1.3 KB | — |
| generation_config.json | Configuration | 3.9 KB | — |
| model.safetensors.index.fp32.json | Configuration | 117.9 KB | — |
| normalizer.json | Configuration | 52.7 KB | — |
| preprocessor_config.json | Configuration | 340 B | — |
| pytorch_model.bin.index.fp32.json | Configuration | 118.0 KB | — |
| special_tokens_map.json | Configuration | 2.1 KB | — |
| README.md | Documentation | 21.8 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
| merges.txt | Tokenizer | 493.9 KB | — |
| tokenizer.json | Tokenizer | 2.5 MB | — |
| tokenizer_config.json | Tokenizer | 282.8 KB | — |
| vocab.json | Tokenizer | 1.0 MB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 24.7 GB
Download from OpenAI
Released by OpenAI through its official repository on Hugging Face. Read the license.
Built From
- Described by arXiv:2212.04356
Evaluations
Each result is shown as reported, with the conditions its reporter stated. None is a SAVRN measurement. A comparison lines two results up only when their configuration, unit and setup are all stated and identical.
| Benchmark | Conditions | Result | Reported by | Revision | Date |
|---|---|---|---|---|---|
| ARTPARK-IISc/Vaani-Benchmark-V1.0 | Task Hindi_WERMetric Hindi_WERComparison conditions not established | 26.8 | Not named Reported by a third party |
Evaluated revision not stated | 2026-06-26 |
| hf-audio/open-asr-leaderboard | Task ami_werMetric ami_werComparison conditions not established | 15.95 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task earnings22_werMetric earnings22_werComparison conditions not established | 11.29 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task gigaspeech_werMetric gigaspeech_werComparison conditions not established | 10.02 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task librispeech_clean_werMetric librispeech_clean_werComparison conditions not established | 2.01 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task librispeech_other_werMetric librispeech_other_werComparison conditions not established | 3.91 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task mean_werMetric mean_werComparison conditions not established | 7.44 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task rtfxMetric rtfxComparison conditions not established | 145.51 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task spgispeech_werMetric spgispeech_werComparison conditions not established | 2.94 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task tedlium_werMetric tedlium_werComparison conditions not established | 3.86 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
| hf-audio/open-asr-leaderboard | Task voxpopuli_werMetric voxpopuli_werComparison conditions not established | 9.54 | open-asr-leaderboard Reported by a third party |
Evaluated revision not stated | 2023-11-07 |
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 24.7 GB |
| 16-bit | 3.1 GB |
| 8-bit | 1.5 GB |
| 4-bit | 0.8 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Built on This Model
- Derived fromwhisper-large-v3-turbo
- Derived fromspeech-emotion-recognition-with-openai-whisper-large-v3
- Derived fromwhisper-large-v3-msp-podcast-emotion-dim
- Derived fromwhisper-large-v3-msp-podcast-emotion
Compare whisper-large-v3
Questions About whisper-large-v3
How much GPU memory does whisper-large-v3 need?
About 3.7 GB at 16-bit and 0.9 GB at 4-bit: the weights (1.5B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run whisper-large-v3 on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Can I use whisper-large-v3 commercially?
Yes. whisper-large-v3 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.
Similar Models
parakeet-ctc-1.1b is an ASR model that transcribes speech in lower case English alphabet. This model is jointly developed by NVIDIA NeMo and Suno.ai teams. It is an XXL version of FastConformer CTC [1] (around 1.1B parameters) model. See the model architecture section and NeMo documentation for complete architecture details. To train, fine-tune or play with the model you will need to install NVIDIA NeMo. We recommend you install it after you've installed latest PyTorch version. There are several ways to use this model. Choose the one that fits your needs. NeMo-Speech.cpp provides a lightweight native C++ runtime for local inference with this model. After installing the runtime: See the…
Quantized MLX weights for beshkenadze/cohere-transcribe-03-2026-mlx-fp16. - model.safetensors - config.json - tokenizer.model - tokenizerconfig.json - preprocessorconfig.json - specialtokensmap.json - keymap.json - conversionsummary.json This checkpoint has been re-validated against the current Swift and Python MLX runtimes. Verified semantic parity on an English fixture: - official CUDA reference path (transformers native Cohere ASR) Fastest and smallest, but introduces a lexical regression on the repo sample (Kaldi → Khaldi). - Generated from the Swift-compatible fp16 checkpoint beshkenadze/cohere-transcribe-03-2026-mlx-fp16. - This repository contains inference artifacts only. Refer to…
Quantized MLX weights for beshkenadze/cohere-transcribe-03-2026-mlx-fp16. - model.safetensors - config.json - tokenizer.model - tokenizerconfig.json - preprocessorconfig.json - specialtokensmap.json - keymap.json - conversionsummary.json This checkpoint has been re-validated against the current Swift and Python MLX runtimes. Verified semantic parity on an English fixture: - official CUDA reference path (transformers native Cohere ASR) Matches fp16 on the repo sample while reducing memory substantially. - Generated from the Swift-compatible fp16 checkpoint beshkenadze/cohere-transcribe-03-2026-mlx-fp16. - This repository contains inference artifacts only. Refer to the upstream Cohere model…
This model is finetuned on top of feature extractor XLS-R from Facebook/Meta. The finetuned model achieves the following results on the test set with a 5-gram KenLM. The numbers in parentheses are the results without the language model: This is one of several Wav2Vec-models our team created during the hosted Robust Speech Event. This is the complete list of our models and their final scores: In parallel with the event, the team also converted the Norwegian Parliamentary Speech Corpus (NPSC) to the NbAiLab/NPSC in Dataset format and used that as the main source for training. We have released all the code developed during the event so that the Norwegian NLP community can build upon it when…