zen3-audio-fast

High-Throughput Transcription

An 809M parameter speech-to-text model optimized for maximum throughput. Delivers fast, accurate transcription for high-volume workloads where processing speed and cost efficiency are the primary requirements.

Specifications

Capabilities

• High-throughput batch audio transcription
• Faster-than-realtime processing
• Multi-language support (100+)
• Timestamp generation (word and segment level)
• Punctuation and capitalization restoration
• Cost-optimized for large-scale transcription pipelines

API Usage

curl https://api.hanzo.ai/v1/audio/transcriptions \
  -H "Authorization: Bearer $HANZO_API_KEY" \
  -F model=zen3-audio-fast \
  -F file=@interview.mp3 \
  -F language=en \
  -F response_format=json

from hanzoai import Hanzo

client = Hanzo(api_key="hk-your-api-key")

with open("interview.mp3", "rb") as audio_file:
    response = client.audio.transcriptions.create(
        model="zen3-audio-fast",
        file=audio_file,
        language="en",
        response_format="json",
    )

print(response.text)

Try It

Open in Hanzo Chat

Resources

• Audio API -- Endpoint documentation
• Technical Report

See Also

• zen3-audio -- Best quality transcription (1.5B)
• zen3-asr -- Real-time streaming speech recognition
• Pricing -- Full pricing table