OpenAI releases transcription AI 'GPT Transcribe' and 'GPT Live Transcribe'
OpenAI has released two AI transcription tools, ' GPT Transcribe ' and ' GPT Live Transcribe .' Both models are touted as being able to understand context even in audio containing technical terms, and to perform highly accurate transcription even in noisy audio.
Transcription | OpenAI API
We're introducing two new transcription models in the API:
— OpenAI Developers (@OpenAIDevs) July 28, 2026
• GPT-Live-Transcribe: built for low-latency live transcription.
• GPT-Transcribe: optimized for asynchronous transcription of completed audio files and batch workloads.
Both models better understand context and… pic.twitter.com/T4rQpKaGmk
GPT Transcribe is an AI-powered transcription tool for recorded audio files. It uses a time-based billing system, with a charge of $0.0045 (approximately 0.74 yen) per minute.
GPT Transcribe Model | OpenAI API
https://developers.openai.com/api/docs/models/gpt-transcribe
The benchmark results for GPT Transcribe are shown below. It has a lower error rate than whisper-1 and is capable of high-precision transcription.
GPT Live Transcribe is an AI transcription tool specializing in real-time processing, such as generating subtitles while someone is speaking. It is more expensive than GPT Transcribe, costing $0.017 (approximately 2.78 yen) per minute.
GPT Live Transcribe Model | OpenAI API
The benchmark results are as follows. The error rate is lower than that of gpt-realtime-whisper, which was released in May 2026.
Both models also boast excellent noise immunity, and OpenAI has demonstrated an example of 'real-time transcription of speech while playing a musical instrument.'
Related Posts:
in AI, Posted by log1o_hf