OpenAI to Launch New Audio Language Model and Hardware by 2027

OpenAI has announced plans to unveil a new audio language model in the first quarter of 2026, which will pave the way for the development of an audio-based hardware device anticipated for 2027. This information comes from a report by The Information, citing sources familiar with the company’s strategic direction.

The report highlights that OpenAI is consolidating various teams across engineering, product, and research to enhance its audio models. Current and former employees indicate that these audio models currently lag behind the text-based models in both accuracy and speed. As a result, there is a pressing need for improvement, particularly since user engagement with the voice interface of ChatGPT has been relatively low, with most users opting for text interactions.

OpenAI’s objective is to significantly enhance its audio capabilities, potentially encouraging users to embrace voice interfaces. Such advancements could lead to broader applications across various devices, including cars. The company envisions a series of physical devices launching in the coming years, beginning with an audio-focused product. Discussions within OpenAI have suggested multiple potential forms for these devices, such as smart speakers and augmented reality glasses, with a clear emphasis on audio functionality rather than screen-based experiences.

The push for audio technology is not unique to OpenAI. Competitors including Google, Meta, and Amazon are increasingly investing in research and development related to voice and audio interfaces. For instance, Meta has recently made strides into the smart glasses market, reflecting a growing trend in the tech industry.

Historically, there have been notable surges in the popularity of voice-activated devices, such as those powered by Alexa and Google Assistant. While these products gained traction among casual technology users, they also faced significant limitations that hindered broader adoption. The new approaches based on large language models, including those that OpenAI is pursuing, may offer innovative solutions and present unique challenges.

Some industry leaders, including former Apple design chief Jony Ive, suggest that voice-controlled devices may be less addictive than traditional screen-based ones. However, they often do not provide substantial evidence to support this assertion. OpenAI’s first audio-centric product is slated for release approximately one year from now, although specific details regarding its design and features remain undisclosed.

As the competition in audio technology intensifies, the coming years will likely reveal how OpenAI and its peers adapt to the evolving landscape of user preferences and technological capabilities.