Spanish speech-to-text API: 2.7% 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`	2.7%	1.1%	20 of 30	0 of 30	0.058× real time
Parakeet TDT 0.6B v3 `nvidia/parakeet-tdt-0.6b-v3`	3.5%	2.0%	17 of 30	0 of 30	0.048× 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	No se trata, sin embargo, de una experimentación, es una prueba que se usa para descartar una o más de las posibles hipótesis, haciendo preguntas y observaciones que guían la investigación científica.
Whisper Large v3	No se trata, sin embargo, de una experimentación. Es una prueba que se usa para descartar una o más de las posibles hipótesis haciendo preguntas y observaciones que guían la investigación científica.
Parakeet (its median clip)	Spoken: La mayor parte de las islas más pequeñas constituyen estados independientes o asociadas con Francia y son populares por sus playas y complejos turísticos de lujo. La mayor parte de las islas más pequeñas constituyen estados independientes o asociadas con Francia, y son populares por sus playas y complejos turísticos de lujo.

Language detection

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

Which model for Spanish

The two models are within the noise of a 30-clip test (2.7% and 3.5%). 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=es" \
  -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="es", 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 Spanish test.

Notes for Spanish

Whisper handles European and Latin American Spanish; a prompt in the variety you want steers regional spelling.

Text-to-speech in Spanish

The reverse direction is covered too: hear the Spanish voices on the Spanish text-to-speech page.

FAQ

How accurate is Whisper for Spanish?

On 30 clean read-speech clips it made 2.7% word errors, about one word in 37. 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 Spanish?

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

Whisper or Parakeet for Spanish?

On the same clips Whisper Large v3 scored 2.7% and Parakeet 3.5%. 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 Spanish speech to English?

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

Related languages: Portuguese (3.4%) · Italian (2.7%) · Catalan (6.8%) · Galician (12%)