OpenAI: GPT Transcribe
Transcriptionopenai/gpt-transcribeGPT Transcribe is OpenAI's speech-to-text model, released on 2026-07-28 alongside gpt-live-transcribe. It delivers lower word error rates than the GPT-4o transcription family across multilingual benchmarks, and is billed by audio duration rather than tokens, keeping costs predictable for transcription workloads. Accessible via the OpenAI-compatible protocol through Ofox.
Released
2026-07-28
Capabilities
Audio Input
Available Providers
Azure
Supported Protocols
openai
Providers
Azure
Input Audio / Hour
$0.27/h
Protocols
openai
/v1/audio/transcriptionsCode Examples
import requestswith open("speech.mp3", "rb") as audio:response = requests.post("https://api.ofox.io/v1/audio/transcriptions",headers={"Authorization": "Bearer YOUR_OFOX_API_KEY"},data={"model": "openai/gpt-transcribe", "language": "en", "provider_options": '{"elevenlabs":{"diarize":true,"tag_audio_events":true,"diarization_threshold":0.2}}'},files={"file": ("speech.mp3", audio, "audio/mpeg")},)response.raise_for_status()print(response.json())
Upload audio to transcribe it into text. Identify speakers, specify the number of speakers and label audio events.
Related Models
Frequently Asked Questions
Simply set your base URL to https://api.ofox.io/v1 and use your Ofox API key. The API is OpenAI-compatible — just change the base URL and API key in your existing code.
OpenAI: GPT Transcribe supports the following capabilities: Audio Input. Access all features through the Ofox.ai unified API.