OpenAI readies audio language model as a step toward audio-first devices

OpenAI plans to announce a new audio language model in the first quarter of 2026, according to a report in The Information.

The model is intended as a deliberate step toward an audio-based physical device. The company has combined teams across engineering, product, and research into a single initiative focused on improving audio models.

Researchers at OpenAI see audio models as lagging behind text models in accuracy and speed. The report also says relatively few ChatGPT users opt for the voice interface, with most preferring text.

Executives hope stronger audio models could shift users toward voice interfaces and allow the technology to be deployed in a wider range of devices, including cars.

OpenAI plans to release a family of physical devices starting with an audio-focused product. Company discussions have included forms such as smart speakers and glasses, with an overall emphasis on audio interfaces rather than screen-based ones.

Key Topics

AI, Audio, Hardware, Openai, Voice, Models