If you are still using ChatGPT’s standard voice or advanced voice, you have about a month left.
OpenAI announced that the two voice modes, Advanced and Standard, will be removed from the voice settings, and those who are still using them will be moved to Live Voice. Behind Live Voice is the GPT-Live model released by OpenAI in July this year.
The official reason is data: after GPT-Live went online, voice usage almost doubled, and now almost everyone who talks to ChatGPT is already using it. It can listen and talk at the same time. It will respond with a "hmm" when you speak, and it will wait quietly when you stop to think about something.
Standard Voice is the earliest version, and the process is very old-school: one model converts your words into text, a second model answers, and a third model reads the answer. The latency is high, but some people just like the sound and rhythm of it. OpenAI originally planned to take it offline in 2025, but later heard that many users said it was very special to them and kept it. This time I really have to leave.
A few details: old voice conversation records will remain in the history; real-time video and screen sharing are still running on Advanced Voice for the time being, and will be switched after GPT-Live supports it.
Another news has a greater impact: starting on December 11, custom GPTs will stop working, and OpenAI will convert them into plug-in form.
For those who have done GPTs, it is best to export and back up your instructions and knowledge base files now. Don't wait until that day to find out that things cannot be opened.