OpenAI has launched three new real-time audio models—GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper—designed to handle complex voice interactions, translation, and transcription. The ...