OpenAI has 3 new AI voice models that the ChatGPT maker says will ‘unlock a new class of voice apps for developers’
OpenAI released GPT-Realtime-2 for reasoning, GPT-Realtime-Translate for 70+ languages, and GPT-Realtime-Whisper for transcription via Realtime API. These enable low-latency voice agents and multilingual apps. Developers gain tools for natural, adaptive conversations.