2026.11.03 Realtime SaaS
Improvements
English Enhanced model:
Improved recognition of a small number of specific terms and phrases. Overall accuracy is unchanged.
Fixes
Standard and Enhanced models:
Realtime sessions are now less likely to become unresponsive when transcription falls behind the audio being sent. A session that falls more than 1 MB behind the audio being sent is now ended with the following
job_errormessage: “Transcription fell too far behind the audio being sent, so the session was ended.”Fixed a case where a valid WAV file could be misidentified as another format during audio preprocessing. Affected files now transcribe correctly.
Updates
Standard and Enhanced models:
Updated Orchestrator Version:
2026.10.24+f85db168f4+17.0.0