Im noticing some random music and sound effects periodically when using the Realtime API.
Ive seen other reports of this, but im curious if others have found a solution to remedy this. Ive tried prompting around it, but to no success.
Hey @Srmuhs which model/voices are you using?
Hi @Sean-Der We are using the gpt-realtime-2025-08-28 model, voice Marin
Hey @Srmuhs I asked Chenguang Zhu and we we removed music background data from post-training.
Can you try the most recent mini model? I believe it is gpt-realtime-mini-2025-12-15
1 Like
Is it possible to use the non-mini model with this update? gpt-realtime-2025-12-15?
@Sean-Der Also if there is a beta for testing, would be happy to some how get involved.