any qery gives response `This model's maximum context length is 16385 tokens. [...]`

The full error reads
`This model's maximum context length is 16385 tokens. However, your messages resulted in 16680 tokens. Please reduce the length of the messages.`

This comes for any request (that previously could be executed successfully).

I tried setting the `max_tokens` to 20000 or some other arbitrary value but to no avail.
Using [chatgpt CLI](https://github.com/kardolus/chatgpt-cli) I'm having no problems with the same queries.

What am I missing?

Reproduce the error in ChatGPT.nvim with a query that previously worked, then inspect the plugin's request handling and compare its token-related request settings with the chatgpt CLI mentioned in the report. Done means previously successful queries no longer fail with the 16385-token context-length error.

Written by the indexing model from the issue text.