whatmodel.sh / stt / MedASR
MedASR
↗ share
+ contribute
MedASR is an open-weights speech-to-text model from Google, released December 2025. Its published size is 0.105B parameters. Licence: health-ai-developer-foundations.
Catalogue identity notes
google/medasr (Health AI Developer Foundations; gated with automatic approval) - an English medical-dictation ASR model; its general-domain LibriSpeech/FLEURS WER is much higher than general models (domain mismatch, not a failed run: 0 empty hypotheses). transcribe.cpp port pins upstream ae1e484 (current main). Repo created 2025-12-18.
Model information sources
Release date ↗ · Parameter count ↗
Hugging Face safetensors metadata total 105,282,816 (current main ae1e484); transcribe.cpp catalog counts 105,316,112 for the converted port.
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 MedASR, out of 176 with a score, then more from Google.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 172 |  wav2vec 2.0 Large 960h | 7±3 | local | ||
| 173 | data2vec Audio Base 960h | 4±4 | local | ||
| 174 | wav2vec 2.0 Base 960h | 2±3 | local | ||
| 175 | MMS-1B-FL102 | -9±5 | local | ||
| 176 | MedASR | -14±4 | local | ||
| more from Google | |||||
| 11 |  Gemini 3.5 Transcribe | 67±6 | api $0.0050/audio min | ||
| 31 | Chirp 3 | 61±2 | api price unknown | ||
| 33 | Gemini 3.5 Transcribe Live | 61±3 | api $0.0090/audio min | ||
see all speech-to-text models · find the ones that fit your machine
$ what it's good at
other tests2 not in a task · 94% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LibriSpeech test-clean WER | 17.9% ✎ | ▃···▁▁▂·▄▄██ beats 0/56 | 67% |
| FLEURS English test WER | 37.5% ✎ | ▃··▁▁▁▁▁▁▃█▁ beats 0/56 | 27% |
$ where to run it
Other ways to run it
- Run it as published · needs ~0.13 GB · from github.com/handy-computer
git clone https://github.com/handy-computer/transcribe.cpp.git && cd transcribe.cpp && cmake -B build && cmake --build build && hf download handy-computer/medasr-gguf medasr-Q8_0.gguf --local-dir models/medasr && build/bin/transcribe-cli --language en -m models/medasr/medasr-Q8_0.gguf audio.wav
Official weights and documentation: huggingface.co/google ↗
$ how the score was worked out · transcribe.cpp model catalog… 94%
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 | 2 | 94% |
| prior · pull toward the middle | 6% | |
| weighted average | = -13.8 |
$ 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