[ ABORT TO HUD ]
SEQ. 1
SEQ. 2

Speech-to-Text (Transcription)

🎙️ Speech & Audio APIs 12 min 200 BASE XP⌨ HANDS-ON LAB

Audio Transcription Models

OpenAI offers multiple transcription models for converting speech to text, from the legacy Whisper to the new GPT-powered models.

Available Models (2026)

ModelQualitySpeedBest For
gpt-4o-mini-transcribeHighest accuracyFastProduction transcription (recommended)
gpt-4o-transcribeVery highMediumComplex audio, heavy accents
whisper-1GoodFastLegacy, basic transcription
const transcription = await openai.audio.transcriptions.create({
  file: fs.createReadStream("meeting.mp3"),
  model: "gpt-4o-mini-transcribe",
  response_format: "verbose_json",  // Includes timestamps
  language: "en"
});
console.log(transcription.text);

Key Features

  • Timestamps: Get word-level or segment-level timing
  • Language Detection: Automatic or manual language specification
  • Translation: Translate non-English audio directly to English text
💡 Pro Tip: Use gpt-4o-mini-transcribe for best results. It significantly outperforms legacy Whisper on noisy audio, accented speech, and alphanumeric content (phone numbers, codes).
⌨ HANDS-ON LABTranscribe Audio from the CLI
⭐ +150 XP

A 40-minute meeting recording just landed in your downloads. Transcribe it with the audio API using multipart form upload - no SDK required.

1Export your OpenAI API key as OPENAI_API_KEY.
2POST meeting.mp3 to /v1/audio/transcriptions as multipart form data (-F) with model gpt-4o-mini-transcribe.
3Re-run with response_format=verbose_json to get timestamps and detected language.
lab-sandbox — simulated environment
INFINITY LAB SANDBOX v2.6 — simulated shell
Type the command for the current objective. Helpers: "hint", "solution", "clear".
$
OBJECTIVE 1 / 3 — type "hint" if stuck
SYNAPSE VERIFICATION
QUERY 1 // 3
Which transcription model is recommended for production in 2026?
whisper-1
gpt-4o-mini-transcribe
gpt-4o-transcribe
speech-to-text-v2