Gemini 2.5 Pro Context Cache Pricing: Per-request vs Cumulative Token Counting?

Gemini API
, ,
1

Hi everyone,

I’m looking for clarification on the Context Cache pricing structure for Gemini 2.5 Pro as outlined in the official documentation.

According to the pricing page, the Context Cache fees are:

  • $0.31 per 1M tokens for prompts ≤ 200k tokens
  • $0.625 per 1M tokens for prompts > 200k tokens
  • $4.50 per 1M tokens per hour (storage price)

My question is about how the 200k token threshold is calculated:

Is the token count calculated per individual API request, or is it cumulative across all requests under the same API key?

For example, if I make multiple requests:

  • Request 1: 150k tokens
  • Request 2: 180k tokens
  • Request 3: 250k tokens

Would Request 3 be charged at the higher rate ($0.625) because it individually exceeds 200k tokens, or would Request 2 also be charged at the higher rate because the cumulative total (150k + 180k = 330k) exceeds 200k?

I want to make sure I understand the billing logic correctly for planning our usage and costs.

Any clarification from the community or Google team would be greatly appreciated!

Thanks in advance!

3

It’s per API request (which includes each message sent as part of a chat conversation flow)

1 Like