OpenAI plans new voice model in early 2026, audio-based hardware in 2027

Voice has lagged in adoption behind screens. OpenAI wants to change that.