Qwen3 ASR 1.7B HF
Qwen · released Jun 26, 2026 · Qwen/Qwen3-ASR-1.7B-hf
Speech to textTranscribes a recording into words
- Type
- Open weightsApache License 2.0
- Languages
- 30
- Size
- 2B
Context measured in tokens
Our take
Written Sep 17, 2026Qwen3 ASR 1.7B HF is a speech-to-text model you can download and run yourself, with a licence that allows commercial use, changes and redistribution. It is accurate on clean read-aloud audio and holds up on podcasts and video, and we list no host for it, so running it yourself is the only route we can point you to.
Use it to transcribe clean read-aloud audio, where it is better than most models on that condition, or to work through a long backlog of recordings, since an hour of audio takes about 4 seconds on the leaderboard's own hardware. It also holds up on podcast, video and accented recordings, where it is better than most models on those conditions. Skip it if you need speaker labelling, timestamps or streaming, none of which these boards measure.
The case for it
- 1.3% of words wrong on clean read-aloud recordings, better than most models on that condition, so clear single-speaker audio is where it earns its place.
- 7.2% of words wrong on podcasts and video, better than most models there, so you are not limited to studio-quality input.
- An hour of audio in about 4 seconds on the leaderboard's own hardware, which makes a long queue of recordings practical to work through.
- You can download it and run it yourself, and the licence allows commercial use, changes and redistribution (Apache License 2.0).
The case against it
- 5.8% of words wrong on accented speech against 1.3% on clean read-aloud recordings, so expect more correction work when speakers are not reading aloud.
- 30 languages are listed, but no source we hold measures accuracy in any language but English, so the other 29 are unverified in our data.
- We list no host for it, so running it yourself is the only route we can point you to.
How good is it?
An open speech-to-text model for turning recordings of speech into written text.
- turning spoken English into written textOpen ASR WER · 10th of 76
- transcribing speakers with a range of accentsAccented speech · 9th of 76
- transcribing podcasts and video audioPodcasts and video · 6th of 92
TranscriptionTurning speech into text4 of 5Open ASR WER · 10th of 76
95.7%
Misses roughly one word in 23, averaged over nine English test sets.
820×35th of 74
an hour of audio in 4 seconds, on the board's own hardware. Your machine will differ.
30
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 ↓ ↑
Chinese · English · Cantonese · Arabic · German · French · Spanish · Portuguese · Indonesian · Italian · Korean · Russian · Thai · Vietnamese · Japanese · Turkish · Hindi · Malay · Dutch · Swedish · Danish · Finnish · Polish · Czech · Filipino · Persian · Greek · Hungarian · Macedonian · Romanian
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. 20.1 GB spare means a 10% error in the size would not change the answer.
GeForce RTX 5090 · 32 GB
Room to spare. 28.1 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. 3.3 GB spare means a 10% error in the size would not change the answer.
These cards answer whether Qwen3 ASR 1.7B HF 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 Qwen3 ASR 1.7B HF
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
- Qwen/Qwen3-ASR-1.7B-hf
- Architecture
- Dense
- Takes in, gives back
- Audio in, text out
- Catalogue slug
- qwen-qwen3-asr-1-7b-hf