OpenAI Releases Three Realtime Audio Models: GPT-Realtime-2, GPT-Realtime-Translate, and GPT-Realtime-Whisper in the Realtime API