OpenAI API adds new voice intelligence features

OpenAI has enhanced its API with new voice intelligence capabilities, enabling developers to build applications for realistic vocal simulation, real-time translation, and live speech-to-text. The new GPT-Realtime-2 model offers GPT-5-class reasoning for complex user requests, while GPT-Realtime-Translate supports over 70 input and 13 output languages. GPT-Realtime-Whisper provides live transcription. These tools aim to transform voice interfaces from simple responses to active conversational agents capable of listening, reasoning, translating, transcribing, and acting. OpenAI has implemented guardrails to prevent misuse for spam or fraud, with conversation halting mechanisms for guideline violations.