OpenAI Introduces Two Voice Transcription AI Models: GPT-Live-Transcribe and GPT-Transcribe with Real-Time and Asynchronous Support

OpenAI has launched two new voice transcription AI models, GPT-Live-Transcribe and GPT-Transcribe, enhancing real-time voice recognition and large-scale audio processing capabilities. These models, available via API, build upon previous real-time voice models like GPT-Live-1 and GPT-Live-1 Mini, expanding OpenAI's voice AI API ecosystem. They offer improved accuracy and contextual understanding compared to existing models, with GPT-Live-Transcribe optimized for low-latency real-time transcription and GPT-Transcribe tailored for asynchronous processing of completed audio files and batch workloads. Applications include video conferencing subtitles, live translation, and customer support, with enhanced noise resistance capabilities. Pricing starts at $0.017 per minute for GPT-Live-Transcribe, while GPT-Transcribe is priced at $4.50 per 1,000 minutes for batch processing.