ChatGPT to Add 8 New Voices: Accents, Styles, and Features Revealed

OpenAI appears to be expanding the voice options available in ChatGPT’s Advanced Voice Mode, with reports suggesting that eight new voices are in development. These new voices, discovered through reverse engineering, offer a variety of genders, accents, and personalities.

New Voice Options Discovered

TestingCatalog uncovered the potential new voice options, and engineer Tibor Blaho shared detailed audio samples of the voices on X. The new voices, reportedly named Fathom, Glimmer, Harp, Maple, Orbit, Rainbow, Reef, Ridge, and Vale, will complement ChatGPT’s existing voices—Ember, Juniper, Cove, and Breeze. Notably, Sky was removed from the lineup following controversy involving actress Scarlett Johansson.

Diverse Accents and Styles

Early reports suggest that the new voices will include different accents and speech patterns. Fathom and Vale are described as having posh British accents, while Reef seems to feature an Australian voice. This diversity is expected to enhance the user experience by providing more options that cater to different regional and personal preferences.

ChatGPT Voice Mode and the Scarlett Johansson Controversy

OpenAI launched ChatGPT Voice Mode in May as part of the GPT-4o model, which integrates text, vision, and audio capabilities into one system. However, the initial launch attracted attention due to comparisons between the demo voice and Scarlett Johansson, who publicly declined OpenAI’s request to use her voice. Despite this, the voice modality was rolled out to a select number of ChatGPT Plus users in July.

Demo Samples and Animal Sound Imitations

Blaho’s demo of the alleged new voices showcased some quirky features. In one passage, the voices read out sentences that included imitations of dog barks and crows. Some of the voices made a simple attempt at “woof” and “caw” sounds, while others tried to mimic the actual animal noises. These imitations received mixed reviews, with some sounding robotic and tinny, though the overall quality of the voices still impressed many.

Uncertain Timeline for Rollout

Currently, ChatGPT’s voice capabilities remain limited to a small group of users, and OpenAI has not provided any further details about when more users will gain access to the audio feature. As the potential new voices are still in the testing phase, their official release date remains unknown.