I'm using Vertex AI's generative model API, and after a request it returns two fields: totalTokens and totalBillableCharacters. According to the Gemini Flash 2.0 Lite pricing page, input tokens cost $0.075 per 1M tokens and output text tokens cost $0.30 per 1M tokens.

However, I'm confused about the relationship between these four terms:

• totalTokens
• totalBillableCharacters (because I send an image as base64 string for analysis)
• Input tokens
• Output text tokens

Specifically:

1. How do totalTokens and totalBillableCharacters relate to input tokens and output text tokens?
2. Given the pricing, how can I calculate the total monetary cost for a request if I only have the API's totalTokens and totalBillableCharacters?

For example, if my API call returns a combined token count, how can I determine the number of input tokens versus output tokens, and then apply the pricing rates to compute the cost?

Any guidance or code examples for converting these token counts to a dollar amount would be appreciated.