[Feature]: Google cost calculation should use correct price when token count is > 200k #369
Description
Feature Type
Provider feature extension
Provider (if applicable)
Description
Some models from Google has a two-tier pricing for input and output tokens depending on the size of the context. E.g. the price per 1M input tokens for Gemini 3 Pro Preview is:
- $2.00, prompts <= 200k tokens
- $4.00, prompts > 200k tokens
Our current implementation does not allow different pricing tiers based on the number of tokens, which doesn't allow the correct cost calculation.
Example Usage
No response
Can you contribute this?
Yes, I can submit a PR