Nyra Health CrisperWhisper
Nyra Health · released Aug 29, 2024 · nyrahealth/CrisperWhisper
Speech to textTranscribes a recording into words
- Type
- Open weightsCreative Commons Attribution-NonCommercial 4.0
- Languages
- 2
- Size
- 1.6B
Context measured in tokens
Our take
Written Sep 4, 2026Nyra Health CrisperWhisper is a compact downloadable speech-to-text model built for meeting-room audio. Its 1.6 billion parameters fit on modest hardware, but the non-commercial licence and thin language coverage limit where it can go.
Pick this for offline, non-commercial meeting transcription you can self-host — it gets fewer words wrong than most on that condition, and processes an hour of audio in about two minutes on reference hardware. Skip it if you need commercial use, more than English or German, or strong performance on podcasts, accented speech or read-aloud material.
The case for it
- Better than most on meeting audio: 7.1% word error rate against a field middle of 10.6%.
- Fast batch processing at 33.3× real time on the leaderboard's hardware.
- Compact 1.6B parameters, suitable for edge or local deployment.
The case against it
- Worse than most on read-aloud, podcasts and video, and accented speech.
- Non-commercial licence only: no production or commercial use permitted, and no hosted providers listed.
- Only English and German supported, and every accuracy figure we hold is English-only.
How good is it?
An open transcription model for turning meeting recordings into text, though it struggles with clean read-aloud audio.
- transcribing recordings of meetings in a roomRecorded meetings · 10th of 92
- transcribing clear recordings of people reading aloudClean read speech · 74th of 92
TranscriptionTurning speech into text
33.3×74th of 74
an hour of audio in 2 minutes, on the board's own hardware. Your machine will differ.
2
Listed on the model card. The accuracy above is English only.
Percentage of words wrong on each set, lower better. Bars are scaled to this model's own worst case; the placing beneath each rate is against every model measured on that set.
Which languages ↓ ↑
German · English
The figures above come from the Open ASR Leaderboard, an independent public test that runs every model on the same recordings. It is the only measurement of transcription quality we know of, so there are no other scores to show.
Each of these is the same transcription job on a different kind of recording, so together they say where it holds up and where it slips — not how closely it follows an instruction.
Every published score for this model7 scoresEvery figure we hold, from 7 boards, with who ran it and a link to the source — including the boards no rating above is built on.
Can you run it yourself?
Comfortable fit
GeForce RTX 4090 · 24 GB
Room to spare. 20.5 GB spare means a 10% error in the size would not change the answer.
Apple M1 Pro (16-core GPU) · 32 GB
Room to spare. 21.7 GB spare means a 10% error in the size would not change the answer.
Apple M2 (8-core GPU, 8GB unified) · 8 GB
Room to spare. 3.7 GB spare means a 10% error in the size would not change the answer.
These cards answer whether Nyra Health CrisperWhisper loads, not how fast it transcribes. Throughput figures for a transcription model come from its text decoder, so treat this as a fit answer rather than a speed one.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first. This is a fit answer: whether it loads, not how fast it transcribes.
When we formed this view
Recent changes
What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
Licence and identifiers
What the licence allowsCreative Commons Attribution-NonCommercial 4.0, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Creative Commons Attribution-NonCommercial 4.0
Weights are downloadable but commercial use is prohibited. Research and personal use only.
Identifiers
- Hugging Face
- nyrahealth/CrisperWhisper
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- nyrahealth-crisperwhisper