Gemini bills thinking tokens on top of output while OpenAI counts them inside, dev's measurements show

qaiser_mehdi · reddit · 2026-09-07

A developer building a cost meter measured how providers count thinking tokens and found three providers with two opposite conventions:

None of the three throws an error if you get it wrong. Caveat: one model, three calls on free-tier access. The verification script is open-sourced and reproducible with a free AI Studio key; the author asks whether other providers use Google's convention.

Original post →

More from Models

Models channel →