I believe OpenAI will be releasing a new whisper model soon and as soon as it’s out Vapi team usually will have it integrated within a few days. Also not sure how 4o will be implemented into Vapi if it’s a multimodal model but I’m sure it’ll have the best STT