There is no ETA or announcement for enabling GPT-4o-Transcribe in Speech Services. By default, Speech recognition utilizes a Universal Language Model as a base model that is trained with Microsoft-owned data and reflects commonly used spoken language.
But you can use Azure OpenAI connectors in your workflows
Feel free to post the idea in feedback forum for speech services.
Thank you.