OpenAI is optimizing its audio artificial intelligence models in preparation for a planned voice-driven personal device. On January 1, according to a report, OpenAI has integrated its engineering, product, and research teams over the past two months to focus on overcoming technical bottlenecks in audio interaction, with the goal of creating a consumer-grade device that can be operated via natural voice commands. Internal researchers at the company believe that the current voice model of ChatGPT lags behind its text model in both accuracy and response speed, and the underlying architectures used by the two are not the same. It is reported that the new voice model will feature more natural emotional expression capabilities and real-time conversation functions, including the ability to handle interruptions in dialogue—a key feature that existing models cannot achieve—and is planned for release in the first quarter of 2026. The report cites informed sources stating that OpenAI also plans to launch a series of screenless devices, including smart glasses and smart speakers, positioning the devices as users' "collaborative companions" rather than mere application portals. However, before launching consumer-grade AI hardware products that support voice commands, OpenAI first needs to change user habits. The team integration focuses on screenless interaction methods. According to the report, OpenAI's current voice model and text model belong to different architectures, resulting in the quality and speed of answers users receive when conversing with ChatGPT via voice being inferior to those from the text model. To solve this problem, OpenAI has completed a key team integration over the past two months. At the organizational level, voice researcher Kundan Kumar, who joined from Character.AI this summer, serves as the core lead for the audio AI project. Product Research Director Ben Newhouse is restructuring the audio AI infrastructure, with multimodal ChatGPT product manager Jackie Shannon also involved. The report cites informed sources stating that the new audio model architecture will be capable of generating more accurate and in-depth responses, supporting real-time conversations with users, and better handling complex scenarios such as dialogue interruptions. In terms of hardware form, OpenAI's judgment aligns with that of Google, Amazon, Meta, and Apple: existing mainstream devices are not suitable for future AI interaction. The OpenAI team hopes users will interact with the device by "speaking" rather than "looking at a screen," believing that voice is the method closest to human instinct for communication. Furthermore, Jony Ive, the former Apple design chief collaborating with OpenAI on hardware development, emphasized that a screenless design is not only more natural but also helps prevent user addiction. In a May interview, he stated: Cultivating user habits emerges as a key challenge. The main obstacle facing OpenAI lies in user behavior. According to the report, most ChatGPT users have not yet developed the habit of voice interaction, with reasons including insufficient quality of the audio model or users being unaware of the feature's existence. To launch an audio-centric AI device, the company first needs to cultivate the habit of users interacting with AI products via voice. Previous reports indicated that OpenAI spent nearly $6.5 billion in early 2025 to acquire io, co-founded by Jony Ive, while simultaneously advancing multiple workstreams including supply chain, industrial design, and model development. The first device is expected to take at least another year to come to market. This timeline implies that OpenAI needs to accumulate a user base and validate the practicality of audio interaction in daily scenarios by improving the existing ChatGPT voice functionality before the product launch.