Tokens and API

Hope all is well!

I have set my LLM to be using gpt3.5, but when I look at my open AI usage charts, I see that both 3.5 and 4 are being used. How do I get it to only use 3.5?

On the charts, I see the term API request. Does that mean that 1 API request = 1 question being asked?

Thank you!

Hi Ben, yes, this is normal. We use ChatGPT-4 Turbo for some specific tasks.

If you have selected ChatGPT-3.5, we will mainly use that model, but some features might need to be executed with a more capable model such as GPT-4 Turbo, so we use it only when needed.

Is it possible to create a bot that becomes a math teacher that can help students explain topics and also generate practice questions for students to try and give feedback? If so, how do you suggest I go about doing it?


Could you provide me with an example of how you would like the bot to prompt users to ask questions?
Are these predefined questions, or are they open-ended, allowing the user to inquire about anything? Please provide us with a brief text example illustrating how you expect the user to pose questions and what responses you anticipate from the chatbot.