Interface GenAiResponseParameters

Response-side GenAI parameters captured for agent invocation telemetry.

interface GenAiResponseParameters {
    cacheReadInputTokens?: number;
    cacheWriteInputTokens?: number;
    finishReasons?: string[];
    inputTokens?: number;
    outputTokens?: number;
}

Properties

cacheReadInputTokens?: number

Number of input tokens read from the cache.

cacheWriteInputTokens?: number

Number of input tokens written into the cache.

finishReasons?: string[]

Finish reasons returned by the model.

inputTokens?: number

Number of input (prompt) tokens consumed by the response.

outputTokens?: number

Number of output (completion) tokens produced by the response.