factor_agent.web.run_view.Tokens#

class Tokens(input, output, cache_hits, reasoning)[source]#

Bases: object

Tokens of LLM calls as the providers report them; a count no reply reported is None.

cache_hits are input tokens the provider read from its cache, and reasoning output tokens the model spent reasoning.

Attributes#

Tokens.input#

Tokens.output#

Tokens.cache_hits#

Tokens.reasoning#