anthropic.internal.AnthOutputTokensDetails
`usage.output_tokens_details`: the breakdown of `output_tokens`. Under extended thinking (this client's own `ThinkingConfig`) `thinking_tokens` is the count the model generated as internal reasoning.
Reference version
Signature
class anthropic.internal.AnthOutputTokensDetailsusage.output_tokens_details: the breakdown of output_tokens. Under
extended thinking (this client's own ThinkingConfig) thinking_tokens is
the count the model generated as internal reasoning.
Source:<builtin>/anthropic/ns_internal/messages.bamlbytes 5286–5346
Fields
thinking_tokens
int | null