Added litellm model config options and improved `_prepare_max_new_tokens` (#967)
* Added litellm model config options and made increase of `max_tokens` for reasoning models more general
* Updated type hint for `timeout` model config option
Co-authored-by: Nathan Habib <30601243+NathanHB@users.noreply.github.com>
* Removed redundant reasoning tag stripping from litellm model
---------
Co-authored-by: Nathan Habib <30601243+NathanHB@users.noreply.github.com>