Higgs Audio v3 STT
Boson AI · released Mar 25, 2026 · bosonai/higgs-audio-v3-stt
Speech to textTranscribes a recording into words
- Type
- Open weightsApache License 2.0
- Languages
- 1
- Size
- 2.7B
Context measured in tokens
Our take
Written Sep 4, 2026Higgs Audio v3 STT is a compact downloadable speech-to-text model from Boson AI that turns English audio into written words. It runs at over 110 times real time and scores well on most measured conditions, though it only handles one language and must be self-hosted.
Pick this for self-hosted English transcription where open weights and a permissive licence matter, or for batch-processing large audio archives quickly. It suits read-aloud audio and financial calls where it scores better than most. Skip it if you need any language other than English, want a hosted provider, or are transcribing heavily accented global English.
The case for it
- Among the more accurate models on most measured English conditions: read-aloud, podcasts, accented speech and meetings.
- Very fast: over 110 times real time, processing an hour of audio in roughly 33 seconds on benchmark hardware.
- Apache 2.0 licence allows commercial use, modification and redistribution.
- Strong on clean speech and financial calls specifically, at 1.07% and 2.07% word error rates.
The case against it
- English only; no measured or supported language coverage beyond one.
- No hosted inference available; you must run it yourself.
- Heavily accented global English lags its other conditions, with more than double the error rate of European-accented speech.
How good is it?
An open speech-to-text model for turning recordings of meetings, podcasts and read-aloud speech into written text.
- turning spoken English into written textOpen ASR WER · 14th of 76
- transcribing recordings of meetings in a roomRecorded meetings · 12th of 92
- transcribing podcasts and video audioPodcasts and video · 14th of 92
- transcribing clear recordings of people reading aloudClean read speech · 13th of 92
TranscriptionTurning speech into text4 of 5Open ASR WER · 14th of 76
95.6%
Misses roughly one word in 23, averaged over nine English test sets.
111×68th of 74
an hour of audio in 32 seconds, 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. 19.6 GB spare means a 10% error in the size would not change the answer.
GeForce RTX 5090 · 32 GB
Room to spare. 27.6 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. 2.8 GB spare means a 10% error in the size would not change the answer.
These cards answer whether Higgs Audio v3 STT 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 Higgs Audio v3 STT
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 allowsApache License 2.0, 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
Apache License 2.0
Fully permissive: commercial use, redistribution, and derivatives allowed. Requires attribution and a copy of the license. Includes an express patent grant.
Identifiers
- Hugging Face
- bosonai/higgs-audio-v3-stt
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- bosonai-higgs-audio-v3-stt