Estonian speech-to-text API: 17% word error, measured

Measured accuracy

Model	Word error rate	Character error rate	Clips with no errors	Empty transcripts	Speed
Whisper Large v3 `openai/whisper-large-v3`	17%	2.4%	4 of 30	0 of 30	0.069× real time
Parakeet TDT 0.6B v3 `nvidia/parakeet-tdt-0.6b-v3`	17%	4.9%	6 of 30	0 of 30	0.049× real time

Fair. Gets the content across, but needs a human pass before it is published or quoted. 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	Springboksi jaoks lõpetas see viiematšilise kaotuste seeria.
Whisper Large v3	Springboksi jaoks lõpetas see viiematsilise kaotuste seeria.
Parakeet (its median clip)	Spoken: Kaugemal põhja pool asuvad suured alad on päris hõredasti asustatud ja kohati peaaegu asustamata metsik loodus. Kaugemal põhjapool asuvad suured alad on päris hõredasti asustatud ja kohati peaaegu asustamata metsik loodus.

Language detection

All 10 clips we sent without a language field were identified as Estonian. Auto-detection is safe here; sending language=et still skips the detection step.

Which model for Estonian

The two models are within the noise of a 30-clip test (17% and 17%). 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.

cURL

curl https://api.inferenceapis.com/v1/audio/transcriptions \
  -H "Authorization: Bearer $INFERENCE_API_KEY" \
  -F "model=openai/whisper-large-v3" \
  -F "language=et" \
  -F "file=@recording.mp3" \
  -F "response_format=verbose_json"

Python · openai SDK

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="et", 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 Estonian test.

FAQ

How accurate is Whisper for Estonian?

On 30 clean read-speech clips it made 17% word errors, about one word in 6. Gets the content across, but needs a human pass before it is published or quoted. Real recordings with noise, accents or several speakers will score worse.

Does language auto-detection work for Estonian?

In our test, yes: all 10 clips sent without a language were identified as Estonian. Sending language=et still saves the detection step on every request.

Whisper or Parakeet for Estonian?

On the same clips Whisper Large v3 scored 17% and Parakeet 17%. 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 Estonian speech to English?

The API returns Estonian text. For English, transcribe first and translate the text with a chat model on the same key.

Related languages: Finnish (6.6%) · Latvian (20%) · Lithuanian (28%)