OpenAI — the creator of the models and products behind ChatGPT — intends to unveil a new audio language model in the first quarter of 2026, and that model is meant as a deliberate step toward an audio-centered physical hardware device, according to a report in The Information.
Drawing on several people familiar with the plans, including current and former staff, The Information says OpenAI has moved to combine teams from engineering, product, and research into a single initiative focused on advancing audio models, which company researchers consider to trail the text models in both accuracy and speed.
The report also notes that relatively few ChatGPT users choose the voice interface, with most preferring text. The expectation is that substantially better audio models could shift users toward voice, enabling the technology to be used in a broader array of devices, such as cars.
OpenAI plans to introduce a line of physical devices in the coming years, beginning with one focused on audio. People inside the company have discussed various forms for future hardware, including smart speakers and glasses, but the emphasis across the range is on audio interfaces rather than screen-based ones.
