Comment by jshreder

Just letting you know it's available right now, just specify `gpt-4o` -- for text streaming anyway. I'd hazard a guess that the audio endpoints are open now, just not documented (like most of the last launches)...