Russian speech-to-text
ExcellentWe sent 30 Russian test clips through this API. Whisper Large v3 made 5.8% word errors, about one word in 17; Parakeet made 7.2%. $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 | 5.8% | 1.2% | 10 of 30 | 0 of 30 | 0.063× real time |
Parakeet TDT 0.6B v3nvidia/parakeet-tdt-0.6b-v3 | 7.2% | 1.6% | 10 of 30 | 0 of 30 | 0.041× 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 | Чем ниже напряжение, тем более позитивна присутствующая в вас жизненная сила. Каждому человеку потенциально дано обрести абсолютные покой и чувство удовлетворения. |
| Whisper Large v3 | Чем ниже напряжение, тем более позитивна присутствующая в вас жизненная сила. Каждому человеку потенциально дано обрести абсолютный покой и чувство удовлетворения. |
| Parakeet (its median clip) | Spoken: Так же, как Луна притягивает Землю, при этом вызывая приливы, так и Млечный Путь воздействует на галактику в Стрельце. Так же как Луна притягивает Землю, при этом вызывая приливы, так и Млечный путь воздействовать на галактику в Стрельце. |
Language detection
Send language=ru. Of 10 clips sent without it, only 5 were identified as Russian; the others were labelled Belarusian (4), English (1). When detection is wrong, Whisper writes the transcript in the language it guessed.
Which model for Russian
The two models are within the noise of a 30-clip test (5.8% and 7.2%). Parakeet is faster; Whisper handles mixed-language audio and has automatic detection across 100 languages. Both cost the same, so try both on a sample of your own recordings. Side by side.
Quickstart
The request is the OpenAI transcription request with a different base URL. language is optional in the API but, as shown above, not optional in practice for Russian.
curl https://api.inferenceapis.com/v1/audio/transcriptions \
-H "Authorization: Bearer $INFERENCE_API_KEY" \
-F "model=openai/whisper-large-v3" \
-F "language=ru" \
-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="ru", 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 Russian test.
FAQ
How accurate is Whisper for Russian?
On 30 clean read-speech clips it made 5.8% word errors, about one word in 17. 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 Russian?
Not reliably. Of 10 clips sent without a language, 5 were identified as Russian; the rest were labelled Belarusian (4), English (1). A wrong label means the transcript is written in the wrong language, so always send language=ru.
Whisper or Parakeet for Russian?
On the same clips Whisper Large v3 scored 5.8% and Parakeet 7.2%. That is within the noise of a 30-clip test; Parakeet is faster, so try it first on long recordings. Both cost $0.002 per minute.
Can it translate Russian speech to English?
The API returns Russian text. For English, transcribe first and translate the text with a chat model on the same key.
Related languages: Ukrainian (6.4%) · Belarusian (47%) · Bulgarian (8.4%)