Why does Vapi charge different prices for different assistants using the same OpenAI model?

it depends on system prompt length as well

Like AI ID mentioned, you could have the same model configuration with two assistants, but if one has a longer prompt it makes the context window larger. It can get even larger by setting the maxTokens value higher, which will result in a higher price if you have a prompt to fit that context window. If you have any questions, feel free to reach out