OpenAI improves its audio models ahead of the launch of a new personal device

OpenAI has reorganized its internal teams to improve audio models ahead of the launch of a new personal device without a screen, with which it intends to focus interaction with AI through voice.

The technology firm has dedicated the last two months to making changes internally with the aim of promoting its audio models, for which it has unified engineering, product and research teams.

This movement aims to modernize its audio models to enable a natural interaction between artificial intelligence and people, with the ability to manage interruptions and speak while speaking, according to The Information.

OpenAI is expected to present a new audio model in early 2026, but work would be aimed at a new family of screenless devices that would act as companions for users.

OpenAI has previously announced its intention to create a device other than a smartphone that will do everything for the user through voice interaction, and that it is developing together with Jonathan Ive, former head of Design at Apple and co-founder of io, a company it acquired last year.

OpenAI CEO Sam Altman explained that they hope to build a device that in addition to doing things for users, is contextually aware of when it should bother them and when it should present them with information or ask for their opinion or not.