Vietnamese speech-to-text
ExcellentWe sent 30 Vietnamese test clips through this API. Whisper Large v3 made 7.9% word errors, about one word in 13. $0.002 per audio minute, word timestamps, SRT and VTT subtitles, files up to 100 MB.
Measured accuracy
| Model | Word error rate | Character error rate | Clips with no errors | Empty transcripts | Speed |
|---|---|---|---|---|---|
Whisper Large v3openai/whisper-large-v3 | 7.9% | 3.8% | 8 of 30 | 0 of 30 | 0.056× real time |
Excellent. Production quality. Transcripts can be used as they are for notes, search and subtitles. The clips are clean, single-speaker read speech, so treat this as a best case for your own recordings. How we measured, and all 83 languages.
What a transcript looks like
The clip with the median error rate from the test, reference text against what the API returned.
| Spoken | Blog là công cụ kích thích sự hợp tác và khuyến khích học sinh mở rộng việc học tập ra ngoài khuôn khổ lớp học truyền thống. |
| Whisper Large v3 | Lo là công cụ kích thích sự hợp tác và khuyến khích học sinh mở gọn việc học tập ra ngoài khuôn khổ lớp học truyền thống. |
Language detection
All 10 clips we sent without a language field were identified as Vietnamese. Auto-detection is safe here; sending language=vi still skips the detection step.
Quickstart
The request is the OpenAI transcription request with a different base URL. language is optional.
curl https://api.inferenceapis.com/v1/audio/transcriptions \
-H "Authorization: Bearer $INFERENCE_API_KEY" \
-F "model=openai/whisper-large-v3" \
-F "language=vi" \
-F "file=@recording.mp3" \
-F "response_format=verbose_json"import os
from openai import OpenAI
client = OpenAI(base_url="https://api.inferenceapis.com/v1", api_key=os.environ["INFERENCE_API_KEY"])
with open("recording.mp3", "rb") as f:
r = client.audio.transcriptions.create(model="openai/whisper-large-v3", file=f, language="vi", response_format="verbose_json")
print(r.text)Subtitles and timestamps
response_format=srt or vtt returns a subtitle file with cues of up to two lines, broken at sentence ends and pauses. verbose_json with timestamp_granularities[]=word returns a start and end time for every word if you want to build your own.
Cost
$0.002 per minute of audio on either model: $0.12 for an hour, $12 for a hundred hours. Duration is measured from the file you upload. There is no daily cap on minutes and no subscription; enough to run your own Vietnamese test.
Notes for Vietnamese
- Tone marks are part of the output; a
promptwith proper nouns helps with names.
FAQ
How accurate is Whisper for Vietnamese?
On 30 clean read-speech clips it made 7.9% word errors, about one word in 13. Production quality. Transcripts can be used as they are for notes, search and subtitles. Real recordings with noise, accents or several speakers will score worse.
Does language auto-detection work for Vietnamese?
In our test, yes: all 10 clips sent without a language were identified as Vietnamese. Sending language=vi still saves the detection step on every request.
Can it translate Vietnamese speech to English?
The API returns Vietnamese text. For English, transcribe first and translate the text with a chat model on the same key.
Related languages: Thai (12%) · Khmer (100%) · Lao (over 100%)