Finnish speech-to-text API: 6.6% 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`	6.6%	1.6%	11 of 30	0 of 30	0.052× real time
Parakeet TDT 0.6B v3 `nvidia/parakeet-tdt-0.6b-v3`	15%	3.3%	6 of 30	0 of 30	0.047× 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	30 prosenttia Machu Picchusta oli ehditty entisöidä vuoteen 1976 mennessä, ja entisöinti jatkuu vielä nytkin.
Whisper Large v3	30 prosenttia machu picchiusta oli ehditty entisöidä vuoteen 1976 mennessä, ja entisöinti jatkuu vielä nytkin.
Parakeet	30 prosenttia Machu Picchuista oli ehdetty entisöidä vuoteen 1976 mennessä, ja entisöinti jatkuu vielä nytkin.

Language detection

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

Which model for Finnish

Whisper Large v3 was more accurate on these clips (6.6% against 15%). Parakeet is faster, which matters mainly on long recordings. Both cost the same. 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=fi" \
  -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="fi", 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 Finnish test.

FAQ

How accurate is Whisper for Finnish?

On 30 clean read-speech clips it made 6.6% word errors, about one word in 15. 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 Finnish?

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

Whisper or Parakeet for Finnish?

On the same clips Whisper Large v3 scored 6.6% and Parakeet 15%. Whisper was more accurate here; Parakeet is faster, which matters mainly for long recordings. Both cost $0.002 per minute.

Can it translate Finnish speech to English?

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

Related languages: Estonian (17%)