Response from calling GenerativeModel.countTokens.
Optional
The breakdown, by modality, of how many tokens are consumed by the prompt.
The total number of tokens counted across all instances from the request.
Response from calling GenerativeModel.countTokens.