Function provider_reports_exclusive_input 

pub fn provider_reports_exclusive_input(provider: &str) -> bool

Returns true when provider reports prompt_tokens exclusive of cache-read and cache-creation tokens.

Anthropic and Minimax (which wraps the Anthropic provider) report prompt_tokens as the count of tokens billed at the full input rate, separate from cache-read and cache-creation tokens. All other providers report prompt_tokens as a total that already includes cached tokens, so no adjustment is needed for them.