Niagara 38m Batch.en
Applied Brain Research · released Feb 19, 2026 · abr-ai/niagara-38m-batch.en
Speech to textTranscribes a recording into words
- Type
- Open weightsCustom licence
- Languages
- 1
- Size
- 38M
Context measured in tokens · download allowed, licence restricts use
Our take
Written Sep 4, 2026Niagara is a 38-million-parameter speech-to-text model built for extreme speed on English audio. It processes an hour of audio in under a second on benchmark hardware, but trades accuracy for that throughput.
Pick this when raw transcription speed is the absolute priority — its throughput is extreme for batch processing. Use it for financial earnings calls, where it performs best, or for edge deployment where model size matters. Skip it if you need accuracy on podcasts, accented speech or meetings, or if you need any language other than English.
The case for it
- Extreme transcription speed: an hour of audio processed in under a second on the benchmark rig.
- Very small footprint at 38 million parameters, fitting almost any hardware.
- Strongest on financial earnings calls, with a lower error rate than its own average.
The case against it
- Worse than most models on every measured condition, from read-aloud to meeting audio.
- Struggles with harder audio: meetings, accented speech and podcasts all sit well above the field middle.
- English only — no other language is measured or supported, and no commercial hosting is available.
How good is it?
An open speech-to-text model for turning recordings into written text, though it trails most models on accuracy.
- turning spoken English into written textOpen ASR WER · 68th of 76
- transcribing recordings of meetings in a roomRecorded meetings · 75th of 92
- transcribing speakers with a range of accentsAccented speech · 67th of 76
- transcribing podcasts and video audioPodcasts and video · 81st of 92
- transcribing clear recordings of people reading aloudClean read speech · 85th of 92
TranscriptionTurning speech into text1 of 5Open ASR WER · 68th of 76
91.3%
Misses roughly one word in 11, averaged over nine English test sets.
4,529×15th of 74
an hour of audio in under a second, on the board's own hardware. Your machine will differ.
1
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.
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 model9 scoresEvery figure we hold, from 9 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. 21.6 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. 22.8 GB spare means a 10% error in the size would not change the answer.
Apple M1 (8-core GPU, 8GB unified) · 8 GB
Room to spare. 4.8 GB spare means a 10% error in the size would not change the answer.
These cards answer whether Niagara 38m Batch.en 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.
Models people weigh against Niagara 38m Batch.en
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 allowsCustom licence, 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
Custom licence
This model ships custom license terms that don't map to a known template. We haven't parsed them, so commercial use, redistribution and derivatives are unverified — review the original terms before shipping.
Identifiers
- Hugging Face
- abr-ai/niagara-38m-batch.en
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- abr-ai-niagara-38m-batch-en