Beingpax
Add OpenAI model provider to the Cloud section
I'm currently using the OpenAI "gpt-4o-mini-transcribe" model, which I successfully added through the custom model section. I found this model works well for me.
It actually seems strange that it is not listed in the Cloud section. It supports three transcribe models:
- gpt-4o-transcribe
- gpt-4o-mini-transcribe
- whisper-1
https://platform.openai.com/docs/guides/speech-to-text
The Whisper-1 model is good, but GPT4O transcribe and GPT4O mini transcribe have some kind of problem. Especially when I was using it. also there were a lot of users talking about it online as well, so I have not added it. Problems like truncating the last part of the transcript and not giving the complete transcript, etc
Hmm, that sounds weird. I haven't faced this issue, but I might not be using it as heavily as you did while developing. Maybe it has evolved since the last time you tried. If you consider it problematic and don't recommend it, that is reasonable from this point.
After using the gpt-4o-mini-transcribe model for a week, I really noticed what you're talking about. Yes, I get those truncations.