Realtime - mismatching event type for audio transcript event

[tbarbugli]

Confirm this is an issue with the Python library and not an underlying OpenAI API

• This is an issue with the Python library

Describe the bug

I receive an event with type response.audio_transcript.done but the ResponseAudioTranscriptDoneEvent type field expects the type to be set to response.output_audio_transcript.done.

# event_data is a dict 
if event_data.type == "response.audio_transcript.done":
   ResponseAudioTranscriptDoneEvent.model_validate(event)

Raises this exception:

pydantic_core._pydantic_core.ValidationError: 1 validation error for ResponseAudioTranscriptDoneEvent
type
  Input should be 'response.output_audio_transcript.done' [type=literal_error, input_value='response.audio_transcript.done', input_type=str]
    For further information visit https://errors.pydantic.dev/2.12/v/literal_error

Note: I am not 100% sure if this is a problem with the Python SDK or with the Realtime API (there is no response.audio_transcript.done in the documentation

To Reproduce

1. Use the Realtime API
2. Try to parse server events sent over data channel using pydantic model_validate

# event_data is a Python dict coming from the peerconnection data channel
if event_data.type == "response.audio_transcript.done":
   ResponseAudioTranscriptDoneEvent.model_validate(event)

Code snippets

OS

macOS

Python version

Python 3.12.9

Library version

2.5.0