Reasoning Tokens
Intermediate tokens generated by large language models with chain-of-thought reasoning capabilities during internal reasoning before producing a final answer. Even when they are not shown to the user, they consume compute and are counted toward output charges.
Timeline (last 90 days)
Thinking tokens are the tokens a reasoning model consumes for internal reasoning before producing an answer; they are usually not shown directly to the user but are still billed as output tokens.
Fewer thinking tokens improve efficiency on simple tasks but may compress the exploration space on complex reasoning tasks.
Thinking tokens are billed as output tokens just like the final answer, and are usually invisible to users or shown only in collapsed form.
All Facts (3)
Thinking tokens are the tokens a reasoning model consumes for internal reasoning before producing an answer; they are usually not shown directly to the user but are still billed as output tokens.
50%UnverifiedFewer thinking tokens improve efficiency on simple tasks but may compress the exploration space on complex reasoning tasks.
50%UnverifiedThinking tokens are billed as output tokens just like the final answer, and are usually invisible to users or shown only in collapsed form.
50%