AiUsage class

Token usage and latency for a single provider call.

All counts are best-effort: a provider that omits a usage field (or returns an unexpected shape) reports 0 for it. cachedInputTokens is the subset of inputTokens served from the provider's prompt cache (Anthropic cache_read_input_tokens, OpenAI prompt_tokens_details.cached_tokens, Gemini cachedContentTokenCount). requestMs is the wall-clock time of the HTTP request, used to compute generation speed. Instances sum with + so the agent loop can aggregate per-turn usage across a whole interaction.

Constructors

AiUsage({int inputTokens = 0, int outputTokens = 0, int cachedInputTokens = 0, int requestMs = 0})

const

Properties

cachedInputTokens → int

final

hashCode → int

The hash code for this object.

no setter, inherited

inputTokens → int

final

outputTokens → int

final

requestMs → int

final

runtimeType → Type

A representation of the runtime type of the object.

no setter, inherited

totalTokens → int

no setter

Methods

noSuchMethod(Invocation invocation) → dynamic

Invoked when a nonexistent method or property is accessed.

inherited

toString() → String

A string representation of this object.

inherited

Operators

operator +(AiUsage other) → AiUsage

operator ==(Object other) → bool

The equality operator.

inherited

Constants

zero → const AiUsage

The additive identity (all counts zero), used as an accumulator seed.