Skip to content

gpt-realtime-2: transcription delay support #1446

@yaniv-peretz

Description

@yaniv-peretz

Describe the bug

Controls how long the model waits before emitting transcription text. Higher values can improve transcription accuracy at the cost of latency. Only supported with gpt-realtime-whisper in GA Realtime sessions.

  • Improving accuracy in non EN languages is a current pain point - please work out to add this feature.

gpt-realtime-2 - transcription delay

Relevant log output

No response

Describe your environment

Minimal reproducible example

No response

Additional information

New feature - not a bug - this a new model capability

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions