Gemini 4 Carbon and Argon double on 1 December
Published 18 November 2026 · figures read from ai.google.dev on the same day
Google's paid-tier table for both models now carries two prices. Input is $2.00 per million tokens through 30 November 2026 and $4.00 from 1 December 2026; output goes from $10.00 to $20.00. Context caching follows, $0.10 to $0.20.
It is not a small print change. For a workload that stays exactly the same, the invoice doubles.
One workload, two invoices
2,500 requests a month, 2,400 input tokens and 600 output tokens each, no cache, standard tier:
| Line | Through 30 Nov | From 1 Dec | Change |
|---|---|---|---|
| Input tokens | $12.00 | $24.00 | +100% |
| Output tokens | $15.00 | $30.00 | +100% |
| Total / month | $27.00 | $54.00 | +$27.00 |
Caching and batch do not cushion it — both are percentages of a number that doubled:
| Variant of the same workload | Through 30 Nov | From 1 Dec |
|---|---|---|
| Standard | $27.00 | $54.00 |
| Half the input read from cache | $21.30 | $42.60 |
| Batch rates | $13.50 | $27.00 |
Second date in the same table
Gemini 3.8 Flash is on its own schedule: $0.75 / $3.75 through 31 December 2026, then $1.50 / $7.50 from 1 January 2027. The same 2,500-request workload goes from $10.13 to $20.25 a month — the same doubling, five weeks later.
What to do with the last week of November
- Re-price pinned commitments now. A quote, a pricing page or a customer contract built on $2.00 / $10.00 is wrong from 1 December.
- Measure whether the model is still the right one. At the new rates, Gemini 4 Carbon sits level with GPT-6.2 Sol and Claude Sonnet 5.5 at $2.00 / $10.00 — but those two are not changing on 1 December.
- Check where the cost actually sits. Output is half the bill in this example. Long answers at 10¢ per thousand tokens are the lever, not prompt trimming.
Price your own workload in the calculator — it uses the rate in force today and flags each model whose price is scheduled to change.
Source
https://ai.google.dev/gemini-api/docs/pricing — read 18 November 2026. Quote: "Input price … $2.00 through November 30, 2026. $4.00 starting December 1, 2026."
Questions this note answers
- When does the Gemini 4 price change take effect?
- Input is $2.00 per million tokens through 30 November 2026 and $4.00 from 1 December 2026; output moves from $10.00 to $20.00 and context caching from $0.10 to $0.20. Gemini 4 Argon changes on the same date.
- What does the increase do to a monthly bill?
- A 2,500-request workload at 2,400 input and 600 output tokens per request, no cache, goes from $27.00 to $54.00 — input $12.00 to $24.00 and output $15.00 to $30.00. Reading half the input from cache makes it $21.30 to $42.60; batch rates make it $13.50 to $27.00.
- Does Gemini 3.8 Flash change at the same time?
- No. Gemini 3.8 Flash stays at $0.75 / $3.75 through 31 December 2026 and moves to $1.50 / $7.50 on 1 January 2027; the same 2,500-request workload goes from $10.13 to $20.25 a month.