Do thinkingBudget tokens count toward billed output in Gemini 2.5 Flash?

Gemini API
,
1

Hi everyone,

I’ve been looking into the billing of Gemini 2.5 Flash, and the output tokens bill makes me curious, so I have a few questions, maybe I misread the documentation. The documentation says this:

Output price (including thinking tokens) - $2.50 per million tokens in USD

Questions

  1. When thinkingBudget > 0, are thinking tokens always accumulated to outputTokens for billing purposes or the API doesn’t charge us for using thinking tokens?
  2. If I set thinkingBudget = 0 (i.e. disable thinking), do I still pay the same $2.50 /M rate for the remaining output tokens?
  3. Is there any scenario where thinking tokens are free or billed at a different rate?

Pricing page for reference: Gemini Developer API Pricing  |  Gemini API  |  Google AI for Developers

Thanks in advance for clarifying!

— Allan

1 Like
2

If the API sends you tokens, you pay for them, at the model specific ‘output’ cost. It doesn’t care where they come from (thinking or response). There’s no ‘thinking token’ cost, it’s all just an output cost, because they’re all just tokens.

2 Likes