Breaking model
OpenAI will remove the legacy audio and realtime model families (gpt-realtime, gpt-audio, gpt-4o-audio, gpt-4o-realtime and their mini variants) and gpt-4o-mini-transcribe-2025-03-20 from the API on January 20, 2027.
What changes
Announced July 20, 2026. The deprecations page lists these as model families or snapshots with a January 20, 2027 shutdown. Replacements: gpt-realtime and gpt-4o-realtime go to gpt-realtime-2.1; gpt-realtime-mini and gpt-4o-mini-realtime to gpt-realtime-2.1-mini; gpt-audio, gpt-audio-mini, gpt-4o-audio and gpt-4o-mini-audio to gpt-audio-1.5; gpt-4o-mini-transcribe-2025-03-20 to gpt-4o-mini-transcribe-2025-12-15.
Breaking: Calls fail after this date.
What to do
Move realtime sessions to gpt-realtime-2.1 (or gpt-realtime-2.1-mini for the mini models) and audio calls to gpt-audio-1.5. Re-pin gpt-4o-mini-transcribe-2025-03-20 to gpt-4o-mini-transcribe-2025-12-15, but note that gpt-4o-mini-transcribe itself shuts down on February 26, 2027. Check realtime WebSocket URLs (?model=…) as well as SDK calls.
Strings to search for
gpt-realtimegpt-realtime-2025-08-28gpt-audiogpt-audio-2025-08-28gpt-realtime-minigpt-audio-minigpt-4o-realtimegpt-4o-audiogpt-4o-mini-realtimegpt-4o-mini-audiogpt-4o-mini-transcribe-2025-03-20