OpenAI has launched three new real-time voice models, signaling a major strategic pivot toward making voice the core interface for artificial intelligence.
The release includes GPT-Realtime-2, which features GPT-5-class reasoning capabilities. It also introduces GPT-Realtime-Translate for live language translation and GPT-Realtime-Whisper for streaming speech-to-text transcription. The technology enables AI to listen, reason, translate, and act within live conversations.
This move directly targets the gap between widespread consumer use of voice AI and its limited business adoption. It positions OpenAI to dominate the next phase of human-AI interaction.