whatmodel.sh / stt / Whisper base
Whisper base
↗ share
+ contribute
Whisper base is an open-weights speech-to-text model from OpenAI, released September 2022. Its published size is 0.074B parameters. Licence: apache-2.0.
Catalogue identity notes
Multilingual base checkpoint (openai/whisper-base@e37978b). Hugging Face mirror created 2022-09-26 with the Whisper release (arXiv 2212.04356). Multilingual checkpoint (openai-whisper name 'base'), distinct from the English-only base.en. Params from the Whisper README size table.
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 Whisper base, out of 176 with a score, then more from OpenAI.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 161 | Whisper tiny.en | 27±3 | local | ||
| 162 |  MMS-1B-all | 25±3 | local | ||
| 163 | Whisper base | 25±3 | local | ||
| 164 | wav2vec 2.0 Large 960h LV-60 self-training | 16±3 | local | ||
| 165 | HuBERT X-Large (LibriSpeech 960h ft) | 16±4 | local | ||
| more from OpenAI | |||||
| 13 | GPT Transcribe | 64±4 | api $0.0045/audio min | ||
| 48 | GPT-Realtime-Whisper | 59±3 | api $0.017/audio min | ||
| 61 | Whisper large-v3-turbo | 57±1 | api $0.0005/audio min +local | ||
see all speech-to-text models · find the ones that fit your machine
$ what it's good at
other tests9 not in a task · 97% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LibriSpeech test-clean WER +1 | 5.1% ✎ | ▁···▁▁▃·▄▄██ beats 3/56 | 33% |
| FLEURS English test WER | 9.9% ✎ | ▁··▁▁▁▃▁▁▃█▁ beats 6/56 | 13% |
| FLEURS Mandarin test error rate | 36.2% ✎ | ▂▃··▂▄▂▂·▆██ beats 1/17 | 12% |
| PESTE · Persian FLEURS (fa_ir) corpus CER | 45.9% ✎ | ▄··▄·▄▄▄··██ beats 1/8 | 12% |
| Danish ASR Leaderboard · mean WER over 5 Danish test sets (RyeAI) | 96.7% ✎ | ▂▃···▃▂▅█▃▂▂ beats 1/16 | 9.5% |
| Betthupferl (German dialect ASR benchmark) · Standard German audio → Standard German transcript WER | 33.8% ✎ | ▄·█·▄▄▄▄·█▄▄ beats 1/10 | 7.3% |
| TurkASR-Bench (7 Turkish sources, 5,200 utterances) · corpus WER | 34.5% ✎ | ▃··▃▃··▅··█▅ beats 1/9 | 5.9% |
| Betthupferl (German dialect ASR benchmark) · dialect audio → Standard German transcript WER | 64.1% ✎ | ▃█····▃▅▃·▅▃ beats 3/10 | 3.2% |
1 more test
| test | result | vs the others in its run | counts |
|---|---|---|---|
| TED-LIUM test WER (short-form) | 13.3% ✎ | ▂····▂▂▂█▃▃▂ beats 9/13 | 1.8% |
$ where to run it
Other ways to run it
version
- Run it as published · needs ~0.7 GB · from github.com/openai (official)
uvx --from openai-whisper whisper audio.mp3 --model base
Official weights and documentation: huggingface.co/openai ↗
$ how the score was worked out · transcribe.cpp model catalog… 56%, PESTE · Persian Speech to Tex… 12%, German dialect ASR Benchmark… 10% and 3 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
| source | results | share of the score |
|---|---|---|
 transcribe.cpp model catalog · LibriSpeech test-clean WER of validated full-precision ports | 3 | 56% |
| PESTE · Persian Speech to Text benchmark (fleurs-fa-ir-v1) | 1 | 12% |
 German dialect ASR Benchmark (Betthupferl; kamaal-08/bavarian-asr-leaderboard) | 2 | 10% |
| Danish Speech-to-Text Leaderboard (RyeAI) · results dataset | 1 | 9% |
| Turkish ASR Leaderboard (TurkASR-Bench, yagmurtuncer) | 1 | 6% |
 Speech Robust Bench (arXiv:2403.07937v3) · clean-speech WER of 25+ open ASR models | 2 | 3% |
| prior · pull toward the middle | 3% | |
| weighted average | = 24.7 |
$ improve this page
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