TTS with a question... the audio back has the answer

OpenAI_Support:

Meanwhile, we suggest you all try using gpt-realtime-1.5 for TTS usage.

@OpenAI_Support The suggested model gpt-realtime-1.5 does not appear to be a valid model for TTS usage.

It also seems to be incorrectly listed in the deprecations documentation as a valid replacement. Additionally, it was already reported here that gpt-4o-mini-tts-2025-12-15 has unresolved issues, so it may not be a suitable replacement either.

If possible, I’d be grateful if this could be forwarded to the team. For now, gpt-4o-mini-tts-2025-03-20 would still be a useful model for the developer community.