Hubert Large Ls960 ft
Meta · released Mar 2, 2022 · facebook/hubert-large-ls960-ft
Speech to textTranscribes a recording into words
- Type
- Open weightsApache License 2.0
- Languages
- 1
- Size
- 0.3B
Context measured in tokens
Our take
Written Sep 4, 2026HuBERT Large Ls960 ft is a tiny downloadable speech-to-text model from Meta that turns English audio into written words at extreme speed. Its Apache licence and 0.3-billion-parameter size suit constrained self-hosted setups, but accuracy lags behind most alternatives on every recording type we track.
Pick this when raw transcription speed is the overriding priority — it processes an hour of audio in about a second on benchmark hardware. Use it for lightweight local deployment where licensing flexibility matters, or environments too constrained for larger models. Skip it if accuracy matters on podcasts, video or meetings, or if you need languages beyond English.
The case for it
- Processes audio at 3,024 times real time on the benchmark rig — among the fastest we track.
- Apache 2.0 licence allows commercial use, modification and redistribution without restriction.
- 0.32 billion parameters fit heavily constrained hardware where larger speech models will not run.
The case against it
- Poor accuracy on most real-world audio: roughly one word in six wrong on podcasts and video, worse than most of 74 models tracked.
- Even on clean read-aloud speech it trails the field, with a higher error rate than the median model.
- English only; no measured data for other languages.
How good is it?
An open speech-to-text model for turning recordings into text, though it trails on meeting and podcast audio.
- transcribing recordings of meetings in a roomRecorded meetings · 90th of 92
- transcribing podcasts and video audioPodcasts and video · 90th of 92
TranscriptionTurning speech into text
3,024×21st of 74
an hour of audio in 1 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 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. 21.4 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. 22.6 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. 4.6 GB spare means a 10% error in the size would not change the answer.
These cards answer whether Hubert Large Ls960 ft 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 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
- facebook/hubert-large-ls960-ft
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- facebook-hubert-large-ls960-ft