whatmodel.sh / stt / Qwen-Audio-3.0-ASR

Qwen-Audio-3.0-ASR

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Qwen-Audio-3.0-ASR is a closed-weights speech-to-text model from Alibaba, released September 2026.

Catalogue identity notes

Non-streaming Qwen-Audio-3.0-ASR (MoE LLM-ASR, successor to Fun-ASR) as evaluated in its technical report; distinct from Qwen-Audio-3.0-ASR-Streaming and from the TTS model qwen-audio-3-0-tts-plus. No API build id or weights.

Model information sources

Release date ↗

61.9±4.2comparison only: no current run option verifiedfrom 2 benchmark results across 1 source* thinly tested, treat the rank as provisional

Score within the Open ASR English short-form WER (March 2026) comparison set only.

$ how it compares

The speech-to-text models ranked either side of Qwen-Audio-3.0-ASR, out of 176 with a score, then more from Alibaba.

#modelscorewhereRAMreleased
22Fish Audio Transcribe 1 Fish Audio62±5api $0.0060/audio min
23Inworld STT-1 Inworld62±6api $0.0025/audio min
24Qwen-Audio-3.0-ASR Alibaba62±4comparison only2026-09
25Cohere Transcribe (03-2026) Cohere62±2api $0.0042/audio min +local~5.5GB2026-03
26Granite Speech 4.1 2B NAR IBM62±4local~2.5GB2026-03
more from Alibaba
44Fun-ASR-Flash Alibaba59±4comparison only

see all speech-to-text models · find the ones that fit your machine

$ what it's good at

other tests2 not in a task · 93% of its score
testresultvs the others in its rundifficultycountswhy
Qwen-Audio-3.0-ASR report · Chinese 4-set CER mean (own API runs) lab2.4% ✎▄········▄·█ beats 3/3avg 6.7%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 55%sharp
Qwen-Audio-3.0-ASR report · English 3-set WER mean (own API runs) lab2.5% ✎█···█··█···█ beats 3/3avg 5.6%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 38%typical

$ where to run it

No verified API or download recorded yet · add one

$ how the score was worked out · Qwen-Audio-3.0-ASR Technical… 93%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Qwen-Audio-3.0-ASR Technical Report (arXiv:2609.07549v2) · lab-reported2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 93%CC BY 4.0
prior · pull toward the middle▇▇ 7%
weighted average= 61.9

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json