Papan peringkat suara-ke-teks bahasa Mongolia

Bandingkan kadar ralat perkataan, kelajuan dan harga antara penyedia suara-ke-teks untuk bahasa Mongolia. Setiap nombor diukur pada audio Mongolia sebenar — bukan pemasaran.

Set data yang digunakan

Korpus pertuturan awam yang digunakan bagi setiap larian penyedia. Buka URL sumber untuk melihat kad set data, lesen dan muat turunnya.

Common Voice 24 (MN)›https://huggingface.co/datasets/btsee/common-voices-24-mn
Shunya Labs Mongolian Speech›https://huggingface.co/datasets/shunyalabs/mongolian-speech-dataset
Common Voice 20 (MN)›https://huggingface.co/datasets/warmestman/common-voice-20-mn-normalized
Modern Voice›Dirakam sendiri untuk penanda aras ini — audio tidak diedarkan, metrik sahaja.

WER vs kelajuan

Setiap titik ialah satu penyedia. Titik ideal di kanan bawah (ralat kecil, laju). Diukur pada audio Mongolia sebenar.

Google STTDuudlaga FlowSpeechmaticsChimegeOpenAI WhisperGPT-4o TranscribeAzure SpeechGladiaGemma 4Whisper large-v3Whisper large-v3-turboWhisper large-v2 MNSeamlessM4T v2MMS-1B-allMoonshine MNwav2vec2 XLSR-53 MNWhisper turbo MNWhisper medium MNElevenLabs Scribe v1ElevenLabs Scribe v2Qwen3-ASR-FlashDolphin smallOmniASR CTC-1BOmniASR LLM-1BVibeVoice ASRWhisper large-v3 MN

Perbandingan penyedia

Isih mana-mana lajur. Klik Butiran untuk pecahan per sampel, harga dan metodologi penyedia.

Sumber terbuka

Penyedia	WER ↑	CER	Ketepatan	Kelajuan	Kependaman	Harga / 1000 minit
Speechmatics	14.4%	7.4%	85.6%	0.51×	13.8s	$8.5	Butiran
Duudlaga FlowMilik kami	14.6%	7.6%	85.4%	1.47×	5.2s	$19.4	Butiran
Chimege	14.6%	6.6%	85.4%	1.51×	4.7s	$11.1	Butiran
Google STT	16.9%	7.4%	83.1%	1.98×	3.7s	$16	Butiran
Whisper large-v2 MN	20.7%	8.3%	79.3%	0.53×	13.5s	$0	Butiran
Azure Speech	20.8%	9.7%	79.2%	0.39×	17.3s	$16.7	Butiran
Whisper medium MNPertindihan	22.5%	8.9%	77.5%	2.7×	2.7s	$0	Butiran
SeamlessM4T v2	24%	11.1%	76%	0.39×	26.3s	$0	Butiran
Whisper turbo MNPertindihan	25.7%	9.2%	74.3%	2.43×	3.0s	$0	Butiran
ElevenLabs Scribe v1	27.1%	10.8%	72.9%	2.14×	3.3s	$6.7	Butiran
ElevenLabs Scribe v2	27.1%	10.6%	72.9%	1.51×	4.7s	$3.7	Butiran
Gemma 4	30.2%	14.7%	69.8%	1×	7.3s	$0	Butiran
OmniASR LLM-1B	33.8%	14.2%	66.2%	0.27×	26.6s	$0	Butiran
Whisper large-v3 MNPertindihan	39.4%	14.9%	60.6%	1.57×	4.5s	$0	Butiran
MMS-1B-all	41.7%	12.3%	58.3%	19.2×	0.4s	$0	Butiran
wav2vec2 XLSR-53 MNPertindihan	45%	16.4%	55%	42.32×	0.2s	$0	Butiran
Dolphin small	49.1%	19.2%	50.9%	1.73×	4.1s	$0	Butiran
OmniASR CTC-1B	51.2%	15.5%	48.8%	0.6×	11.7s	$0	Butiran
GPT-4o Transcribe	53%	27.6%	47%	2.11×	3.7s	$6	Butiran
Moonshine MNPertindihan	58.8%	45.2%	41.2%	8.44×	0.8s	$0	Butiran
Whisper large-v3	89.5%	37.4%	10.5%	1.29×	6.0s	$0	Butiran
Gladia	90.1%	38.9%	9.9%	0.75×	9.4s	$10.2	Butiran
Whisper large-v3-turbo	99%	54.4%	1%	0.91×	8.7s	$0	Butiran
Qwen3-ASR-Flash	103.7%	86.4%	-3.7%	2.19×	3.2s	$2.1	Butiran
OpenAI Whisper	105.2%	63.7%	-5.2%	1.5×	4.5s	$6	Butiran
VibeVoice ASR	107%	61%	-7%	0.53×	12.1s	$0	Butiran
Gemini Flash	Penanda aras akan datang

Semua keputusan diukur pada dataset Mongolia sebenar. Kelajuan ialah × masa nyata (lebih tinggi = lebih laju). WER lebih rendah lebih baik.

Pertindihan— Model dengan lencana ini dilatih pada data yang bertindih dengan korpus penanda aras (Common Voice Mongolia — 113 daripada 173 sampel — atau korpus awam Shunya Labs). Skor pada korpus yang bertindih menjadi tinggi kerana hafalan; nilaikan model ini melalui jadual setiap set data di halaman butirannya, terutamanya korpus yang TIDAK digunakan untuk melatihnya.

Duudlaga Voice Set

92 samples · 10.2 min · 28 systems

A second corpus recorded first-party for this benchmark, covering what the public Mongolian datasets barely contain: modern loanwords, English/Mongolian code-switching, numbers, dates, and commands. Every system is measured on the same 92 recordings.

Penyedia	WER	CER	Ketepatan	Kelajuan
1Duudlaga FlowMilik kami	9.5%	5.3%	90.5%	1.27×
2Google STT	11.7%	5.5%	88.3%	2.28×
3Chimege	14.5%	6.8%	85.5%	1.48×
4Speechmatics	15.3%	7.6%	84.7%	0.5×
5Gemini 3.5 Flash	19.5%	11.4%	80.5%	0.06×
6SeamlessM4T v2Sumber terbuka	22.9%	10.4%	77.1%	4.32×
7Whisper large-v2 MNSumber terbuka	23.2%	10.9%	76.8%	0.55×
8ElevenLabs Scribe v1	23.4%	9.3%	76.6%	2.09×
9ElevenLabs Scribe v2	24.6%	9.8%	75.4%	1.47×
10Azure Speech	26.1%	10.9%	73.9%	0.32×
11Gemma 4Sumber terbuka	27.7%	13.2%	72.3%	1.04×
12OmniASR LLM-1BSumber terbuka	34.5%	14.5%	65.5%	0.27×
13Whisper medium MNSumber terbukaPertindihan	35.1%	14.8%	64.9%	2.71×
14Whisper turbo MNSumber terbukaPertindihan	37.6%	14.1%	62.4%	2.59×
15Whisper large-v3 MNSumber terbukaPertindihan	41%	17.1%	59%	1.58×
16GPT-4o Transcribe	44.3%	23.7%	55.7%	2.62×
17MMS-1B-allSumber terbuka	45.3%	15.1%	54.7%	20.52×
18Dolphin smallSumber terbuka	46.1%	19%	53.9%	1.69×
19W2v-BERT 2.0 MNSumber terbukaPertindihan	46.3%	18.2%	53.7%	28.12×
20OmniASR CTC-1BSumber terbuka	54.9%	18.9%	45.1%	0.59×
21wav2vec2 XLSR-53 MNSumber terbukaPertindihan	56.3%	23.9%	43.7%	45.7×
22Whisper large-v3Sumber terbuka	87.3%	36.6%	12.7%	1.55×
23Gladia	88.1%	39.9%	11.9%	0.74×
24Moonshine MNSumber terbukaPertindihan	96.2%	75.7%	3.8%	8.36×
25Whisper large-v3-turboSumber terbuka	98.1%	57.1%	1.9%	1.19×
26Qwen3-ASR-Flash	103.6%	89.9%	-3.6%	2.07×
27OpenAI Whisper	104.6%	63.8%	-4.6%	1.32×
28VibeVoice ASRSumber terbuka	110.1%	64%	-10.1%	0.4×

Dirakam sendiri untuk penanda aras ini — audio tidak diedarkan, metrik sahaja.