Token calculation and pricing in follow up questions?

Can someone highlight how the tokens are calculated in a follow-up question? Does the response to the previous question become part of the new question?
For e.g
Me => generate a report for XYZ,
ChatGPT => abcdefg
Me => Now add qwerty… to the report
ChatGpt => abcdefg qwerty

When I ask it to add ‘qwerty’ to the report, does my command take add previous response(abcdefg) from chatgpt to the new query, or is some context preserved?
How does the token calculation and therefore pricing work in the second case?