Interface ITokenUsage

Token usage statistics for LLM API calls.

Interface

ITokenUsage

Example

const usage: ITokenUsage = {
  promptTokens: 150,
  completionTokens: 75,
  totalTokens: 225
};
interface ITokenUsage {
    promptTokens: number;
    completionTokens: number;
    totalTokens: number;
    cacheReadTokens?: number;
    cacheCreationTokens?: number;
}

Properties

promptTokens completionTokens totalTokens cacheReadTokens? cacheCreationTokens?

Properties

promptTokens

promptTokens: number

Number of tokens in the prompt (input)

completionTokens

completionTokens: number

Number of tokens in the completion (output)

totalTokens

totalTokens: number

Total tokens used (prompt + completion)

Optional cacheReadTokens

cacheReadTokens?: number

Tokens served from the provider's prompt-prefix cache (Anthropic cache_read_input_tokens). Optional — undefined when no aggregated message reported cache activity, so consumers can distinguish "provider did not report" from "zero hits".

Optional cacheCreationTokens

cacheCreationTokens?: number

Tokens written to the provider's prompt-prefix cache as a new entry (Anthropic cache_creation_input_tokens). Same undefined-vs-zero convention as cacheReadTokens.

Settings

Member Visibility

Theme

On This Page

promptTokenscompletionTokenstotalTokenscacheReadTokenscacheCreationTokens