Qwen3-ASR-Flash — Mongolian Speech-to-Text Benchmark
Real results from the Qwen3-ASR-Flash Batch API (enhanced operating point, language auto (mn rejected)) evaluated across 4 Mongolian datasets and 265 audio samples. Every number below is measured — not marketing. WER, speed, and pricing are shown as-is, brutally honest.
265/265 samples transcribed · 100% success rate
Lower is better · across 265 samples
Pricing
Qwen3-ASR-Flash list pricing for batch transcription. No discounts, no negotiated rates applied — the raw per-minute rate.
Pricing source: Qwen3-ASR-Flash public pricing. Duudlaga Flow is shown for context only — this page isolates Qwen3-ASR-Flash so the number is not padded by our own product.
WER & CER by dataset
Word and character error rates per dataset. Lower is better — and these are the real Qwen3-ASR-Flash numbers, which are weak on Common Voice 24.
Dataset summary
Aggregate accuracy, speed, and timing for each dataset.
| Dataset | Source | Samples | Success | WER | CER | Accuracy | Speed | Audio (s) | Proc (s) |
|---|---|---|---|---|---|---|---|---|---|
| Common Voice 24 (MN) | Hugging Face → | 59 | 59/59 | 105.2% | 90% | -5.2% | 2.81× | 330.6s | 117.7s |
| Shunya Labs Mongolian Speech | Hugging Face → | 60 | 60/60 | 102.3% | 75.2% | -2.3% | 1.94× | 645.1s | 332.9s |
| Common Voice 20 (MN) | Hugging Face → | 54 | 54/54 | 104.5% | 90.4% | -4.5% | 2.71× | 269.8s | 99.5s |
| Modern Voice | — | 92 | 92/92 | 103.2% | 89.1% | -3.2% | 2.04× | 603.8s | 295.3s |
WER vs speed — per sample
Each dot is one audio sample. The sweet spot is the bottom-left (low error, fast). Qwen3-ASR-Flash sits high on error for many Common Voice 24 samples.
Per-sample results
Ground truth shown verbatim in the Expected column. The result column highlights only the words Qwen3-ASR-Flash got wrong, in red — no strikethrough/swap gymnastics, just the mistakes.
| # | Audio | Sample | Dataset | Expected (ground truth) | Qwen3-ASR-Flash result | WER | CER | I/D/S |
|---|---|---|---|---|---|---|---|---|
| 1 | btsee_0001 | Common Voice 24 (MN) | Гэхдээ амьсгал хураахаасаа өмнө танд мэдэж байгаагаа хэлье. | كشدا مسخرة خاصة ومن تندم يدي جاغا هي لي. | 112.5% | 87.9% | 1/0/8 | |
| 2 | btsee_0002 | Common Voice 24 (MN) | Надад заяасан аз жаргал гэдэг ердөө гуравхан сарын хугацаатай байсан гэж үү? | 나타자야상아처럼깨끗이剃도그러곤살에혹자떼버리는게좋. | 100.0% | 100.0% | 0/11/1 | |
| 3 | btsee_0003 | Common Voice 24 (MN) | Одоо бид өөрсдөө өвчин эмгэгээсээ салахыг хичээцгээе. | أعطى بيدروس دوتشين فيكيسي سفيرًا لتشيكيا. | 100.0% | 92.3% | 0/1/6 | |
| 4 | btsee_0004 | Common Voice 24 (MN) | Би бол голдуу хээрээр гэр, хэцээр дэр хийж явдаг хүн. | Við varst kalt og hér er gott, ég þyrfti þér ekki að segja það kom. | 150.0% | 111.8% | 5/0/10 | |
| 5 | btsee_0005 | Common Voice 24 (MN) | Хан хурмаст уурлаж, Болдоггүй Бор өвгөнийг хор луугаараа ниргүүлэхээр явуулжээ. | خانفرم استورت سپورت کویبرا ونیو خیرت ثووگارنیرو شریمچی. | 100.0% | 92.2% | 0/2/8 | |
| 6 | btsee_0006 | Common Voice 24 (MN) | Алив наашаа ороод ир гээд гэртээ оров. | عجبنه شعرة كتكشت تعرف. | 100.0% | 91.9% | 0/3/4 | |
| 7 | btsee_0007 | Common Voice 24 (MN) | Өө өндөр дээдэс таны тухайд би баталж чадахгүй. | أو أنظر تلك التنيطات ببطء شتى. | 100.0% | 89.1% | 0/2/6 | |
| 8 | btsee_0008 | Common Voice 24 (MN) | Харин гурав дахь удаагаас эхлэн хүмүүсийг сонирхож эхлэв. | Herhangi rüptükle takas etmek umutsuzca zor çıktı. | 100.0% | 92.9% | 0/1/7 | |
| 9 | btsee_0009 | Common Voice 24 (MN) | Та нар очингуутаа шөл л өгч үз. | Таны орчингод шүдлэл тохуч. | 100.0% | 60.0% | 0/3/4 | |
| 10 | btsee_0010 | Common Voice 24 (MN) | Ерөөсөө литр үйлдвэрээсээ салаагүй явсан юм чинь. | 이로써 롤드컵에서 사타구니 없앤 첫. | 100.0% | 91.7% | 0/2/5 |
Methodology
How these numbers were produced.
Provider: Qwen3-ASR-Flash (Alibaba qwen3-asr-flash hosted transcription with automatic language detection — Mongolian is not a supported language (the API rejects the 'mn' code).).
Endpoint https://dashscope-intl.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation. Language auto (mn rejected). Operating point qwen3-asr-flash. Diarization none.
Datasets: Common Voice 24 (MN), Shunya Labs Mongolian Speech, Common Voice 20 (MN), Modern Voice — 265 samples, 1849.3s of audio total.
- Common Voice 24 (MN)https://huggingface.co/datasets/btsee/common-voices-24-mn
- Shunya Labs Mongolian Speechhttps://huggingface.co/datasets/shunyalabs/mongolian-speech-dataset
- Common Voice 20 (MN)https://huggingface.co/datasets/warmestman/common-voice-20-mn-normalized
- Modern Voice—
Metrics: WER and CER are computed with a standard word/character Levenshtein alignment, normalized for case and punctuation. Accuracy = 100 − WER. Speed factor = audio duration ÷ processing time (× realtime). All requests are real Qwen3-ASR-Flash Batch API calls, not cached or simulated.
Diff highlighting: The result column aligns to the ground truth and colors every substitution and insertion red. Deletions (words missing from the result) are not shown in the result column — the Expected column already holds the full ground truth as-is.