Audio
114d agoOpenAI Launches Three Real-Time Voice Models
Three new models, including GPT-Realtime-2, now enable real-time reasoning, live transcription, and translation across 70+ languages. OpenAI claims the reasoning capabilities match GPT-5. This update eliminates the latency found in previous voice interfaces.
The Signal
Developers can now integrate low-latency, high-reasoning audio interactions directly into their applications.