* Return a shallow copy of this provider with the per-request completion * budget clamped to `maxCompletionTokens`. Optional because not every * backend benefits from a client-computed cap. * * When `options` are provided, implementations may further tighten the cap * based on their o
(
maxCompletionTokens: number,
options?: MaxCompletionTokensOptions,
)
| 216 | * original. See `KimiChatProvider._clone()` for the rationale. |
| 217 | */ |
| 218 | withMaxCompletionTokens?( |
| 219 | maxCompletionTokens: number, |
| 220 | options?: MaxCompletionTokensOptions, |
| 221 | ): ChatProvider; |
no outgoing calls
no test coverage detected