GPT-4o cost is different for every assistant, for some assistants it is 0.05, for some it is 0.07, for some it is 0.09. Why is that? What are reasoning for these varying prices? Is it VAPI's AB testing, or prompt length, Knowledge base size, etc. determine the price?

Prices differ because Vapi bills you for what you actually use: • The specific GPT-4o variant your assistant is configured with • The actual tokens consumed by prompts and responses • Telephony, STT, and TTS services attached to that assistant Vapi’s base rate is fixed, but the total cost per minute changes depending on how much model usage and ancillary services each assistant consumes. That’s why one assistant might show $0.05/min while another shows $0.07 or $0.09/min.

Doesn't make sense, I was expecting helpful answer not GPT generated response.

Hi @Vengeance The token in and token out are exactly same for those 5 assistances? If so, then price must be same if different, then, price would be different

Here’s a clearer and more professional version you can send: ---- Hi, The pricing is aligned with the official OpenAI model rates. Any difference you may notice is typically due to variations in usage. If you need further assistance, please share your call ID, and we’ll be happy to take a closer look.

thanks, that explains a lot

@Vengeance happy to help