Canary 1b v2
NVIDIA · released Aug 4, 2025 · nvidia/canary-1b-v2
Speech to textTranscribes a recording into words
- Type
- Open weightsCreative Commons Attribution 4.0
- Languages
- 25
- Size
- 1B
Context measured in tokens
Our take
Written Sep 4, 2026Canary 1b v2 is NVIDIA's one-billion-parameter speech-to-text model with a permissive open licence. It transcribes an hour of audio in about two seconds on benchmark hardware, though its accuracy sits below the middle of the pack on most audio conditions we measure.
Pick this when raw throughput matters above all else, or when you need a permissive licence that allows commercial redistribution with attribution. It is also viable for clean read-aloud audio where a 1.8% error rate is acceptable. Skip it if you need measured accuracy on podcasts, meetings or accented speech, or if you want a hosted API rather than self-hosting.
The case for it
- Transcribes at 1,821 times real-time on the benchmark rig — an hour of audio in roughly two seconds.
- CC-BY 4.0 licence allows commercial use and redistribution with attribution.
- 25 languages claimed on the model card, including several less commonly supported European ones.
The case against it
- Below-median accuracy on most measured audio: 6.4% overall against a 5.8% median, and worse-than-most on read-aloud, podcasts, accented speech and meetings.
- No commercial hosting available; you must run it yourself.
- Weak where most users need it: 9.1% on podcasts and video, 13.0% on real meetings, both well below the middle of the field.
How good is it?
An open speech-to-text model for turning recordings into written words, though podcast and video audio is where it struggles most.
- transcribing podcasts and video audioPodcasts and video · 71st of 92
TranscriptionTurning speech into text3 of 5Open ASR WER · 51st of 76
94.3%
Misses roughly one word in 18, averaged over nine English test sets.
1,825×29th of 74
an hour of audio in 2 seconds, on the board's own hardware. Your machine will differ.
25
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 ↓ ↑
Bulgarian · Croatian · Czech · Danish · Dutch · English · Estonian · Finnish · French · German · Greek · Hungarian · Italian · Latvian · Lithuanian · Maltese · Polish · Portuguese · Romanian · Slovak · Slovenian · Spanish · Swedish · Russian · Ukrainian
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. 20.9 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. 22.1 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. 4.1 GB spare means a 10% error in the size would not change the answer.
These cards answer whether Canary 1b v2 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 Canary 1b v2
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 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 4.0
Permissive content license: any use with attribution. Common for datasets and some model weights.
Identifiers
- Hugging Face
- nvidia/canary-1b-v2
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- nvidia-canary-1b-v2