Hacker News Reader

You pay for chain of thought tokens as well. Ex: Imagine asking for a single number that takes 1h to compute

That's included in my estimate. Take for example Opus 5.5: https://openrouter.ai/anthropic/claude-opus-5.5

Currently for today, it's 1.19T input tokens, 5.41B reasoning tokens, and 4.29B completion.

Therefore, reasoning + completion is 0.8% of the total usage. Actually below my estimate!