Interface: LanguageModelUsage

Usage information for a language model call.

If your API return additional usage information, you can add it to the provider metadata under your provider's key.

Properties

Property	Type	Description	Defined in
`inputTokens`	{ `cacheRead`: `number` \| `undefined`; `cacheWrite`: `number` \| `undefined`; `noCache`: `number` \| `undefined`; `total`: `number` \| `undefined`; }	Input token usage breakdown.	packages/protocol/src/language-model/model.ts:134
`inputTokens.cacheRead`	`number` \| `undefined`	Input tokens read from cache.	packages/protocol/src/language-model/model.ts:148
`inputTokens.cacheWrite`	`number` \| `undefined`	Input tokens written to cache.	packages/protocol/src/language-model/model.ts:153
`inputTokens.noCache`	`number` \| `undefined`	Input tokens that were not cached.	packages/protocol/src/language-model/model.ts:143
`inputTokens.total`	`number` \| `undefined`	Total input tokens used.	packages/protocol/src/language-model/model.ts:138
`outputTokens`	{ `reasoning`: `number` \| `undefined`; `text`: `number` \| `undefined`; `total`: `number` \| `undefined`; }	Output token usage breakdown.	packages/protocol/src/language-model/model.ts:159
`outputTokens.reasoning`	`number` \| `undefined`	Reasoning/thinking tokens.	packages/protocol/src/language-model/model.ts:173
`outputTokens.text`	`number` \| `undefined`	Text generation tokens.	packages/protocol/src/language-model/model.ts:168
`outputTokens.total`	`number` \| `undefined`	Total output tokens used.	packages/protocol/src/language-model/model.ts:163