Beginner question on Input Tokens

Hi, I’m currently experimenting with an Assistant using gpt-3.5-turbo-1106. I’ve added a few PDFs (1MB total), started asking some simple questions and my Input token count is out of the roof. Like 42K input tokens for “is Scotch whisky the same as Irish whiskey?”.
The output is way longer and is only 380 tokens. Am I missing something?

Thank you.

The text from the PDFs is being searched for relevant information then it’s automatically and dynamically added to the context as input so the model will be better able to answer the questions asked.

This is normal and expected behavior.

Thank you for the confirmation, appreciated.