OpenAI plans to introduce a new audio model in the first quarter of 2026, which will be a major milestone towards the creation of voice-controlled hardware devices. This was reported by the Information, citing sources familiar with the company’s plans, including current and former employees.
According to the publication, OpenAI has brought together several teams from engineering, product development, and research into a single initiative aimed at improving audio models. The company believes that current audio models lag behind text-based models in accuracy and speed.
It also notes that few ChatGPT users prefer voice interfaces over text. OpenAI hopes that significant improvements to audio models will change user behavior and increase the popularity of voice interfaces, allowing for the implementation of models and products in a wider range of devices, such as cars.
OpenAI plans to release a line of hardware devices, starting with audio-focused devices. Various form factors for future devices are being discussed internally, including smart speakers and glasses, but the emphasis is on voice interfaces rather than screens.
