Audio22h ago
OpenAI Launches Real-Time Voice Models
Three new models, including GPT-Realtime-2, now bring reasoning capabilities matching GPT-5 to live conversations. The suite adds real-time transcription and translation across 70+ languages via GPT-Realtime-Whisper. These updates eliminate the latency found in previous voice iterations. Developers can now build low-latency, high-reasoning voice interfaces for complex user interactions.