OpenAI has finally officially rolled out the Alpha test version of GPT-4o Advanced Voice Mode to some Plus users.
This new feature allows users to interrupt conversations at any time, and the system can also perceive and respond to users' emotional fluctuations. However, the previously anticipated video and screen sharing features will be launched later.
OpenAI plans to gradually expand the number of test users, with all Plus members expected to be able to experience this feature by this fall.
Some overseas users have already gained access to the test and shared their experiences:
-
As a language learning tool:
- GPT-4o can act as a foreign language practice partner, showing patience and attention to detail.
- It can fluently switch between multiple languages, such as French, Spanish, Urdu, etc.
- The Chinese effect still has room for improvement, sounding slightly mechanical.
-
Entertainment functions:
- Able to understand and tell jokes.
- Can narrate horror stories and simulate corresponding sound effects.
- Capable of reciting tongue twisters and performing beatbox.
-
Video understanding function (experienced by a few users):
- Able to understand video content in real-time and interact.
- Can perform real-time video translation.
Currently, GPT-4o offers four voice options: Juniper, Breeze, Cove, and Ember, all from professional voice actors. OpenAI has also developed new filters to prevent the generation of music or audio content that might infringe on copyrights.
Although OpenAI's product release pace is unstable, its technological strength is still widely recognized. Users have high expectations for OpenAI's new features, while also hoping that the company can quickly launch the full version and more innovative products.
[OpenAI Official Notification Example]