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

OpenAI, the innovative force behind ChatGPT, is gearing up to unveil a cutting-edge audio language model in the first quarter of 2026. This development marks a strategic move toward launching an audio-centric physical device by 2027, as reported by The Information. According to sources with inside knowledge, including current and former employees, OpenAI is actively merging various teams from engineering, product development, and research. This collaborative effort aims to enhance audio models, which are currently perceived to be less accurate and slower compared to their text-based counterparts. Interestingly, data indicates that only a small fraction of ChatGPT users engage with the voice interface, with the majority favoring text interactions. The company's ambition is to significantly upgrade its audio models, potentially shifting user preferences toward voice interfaces. This shift could pave the way for broader applications of their technology across diverse devices, particularly in automotive settings. OpenAI envisions a series of physical products in the near future, starting with an audio-focused device. Discussions within the company have explored various formats for these future products, including smart speakers and smart glasses, with a clear focus on enhancing audio interaction rather than traditional screen-based interfaces.