The model thinks before it answers
Reasoning models spend tokens on a private chain of thought before the final answer.
1The model just blurted “12”. Flip the switch to make it think first.your turn
the problem
A drawer holds 12 red socks and 12 blue socks in the dark. How many socks must you pull to be sure you have a matching pair?
✗
the answer
12 sockswrong, pattern-matched, never worked it out
Answered in one shot. The prompt is saturated with 12s, so the first token gets pulled toward 12 before any working-out can happen.
→ continue← backR replay
If thinking is just tokens, you can buy more of it, a new way to scale.
4.9 Test-time compute: pay at answer-timeBuilds on4.1Tuning the model's creativity
Common questions
What is "The model thinks before it answers" about?
Reasoning models spend tokens on a private chain of thought before the final answer.
What problem does it solve?
Hard problems need working-out. But the model answers in a single pass, no scratch paper to think on.
What will I be able to do after this lesson?
You can explain why reasoning models think before answering, and that the 'thinking' is generated tokens.
What comes next?
If thinking is just tokens, you can buy more of it, a new way to scale.