OpenaiCompletionUsageCompletionTokensDetails
OpenaiCompletionUsageCompletionTokensDetails =
object
Defined in: src/client/types.gen.ts:3737
Breakdown of tokens used in a completion.
Properties
accepted_prediction_tokens?
optionalaccepted_prediction_tokens:number
Defined in: src/client/types.gen.ts:3742
When using Predicted Outputs, the number of tokens in the prediction that appeared in the completion.
audio_tokens?
optionalaudio_tokens:number
Defined in: src/client/types.gen.ts:3746
Audio input tokens generated by the model.
reasoning_tokens?
optionalreasoning_tokens:number
Defined in: src/client/types.gen.ts:3750
Tokens generated by the model for reasoning.
rejected_prediction_tokens?
optionalrejected_prediction_tokens:number
Defined in: src/client/types.gen.ts:3757
When using Predicted Outputs, the number of tokens in the prediction that did not appear in the completion. However, like reasoning tokens, these tokens are still counted in the total completion tokens for purposes of billing, output, and context window limits.