Google Gemini · Pricing checked October 19, 2026
Gemini 4 Argon pricing.
Plan beyond the introductory rate.
Standard input and output rates are $2 and $10 per million tokens through November 30, 2026. Google announces $4 and $20 thereafter. Compare the same workload across both periods below.
Monthly token budget calculator
Enter your total monthly usage in millions of tokens. Use uncached input and total billable output, including thinking tokens where applicable.
USD estimate. Excludes cache discounts/storage, batch or priority tiers, tool calls, hosting, taxes and negotiated rates. Retries and additional agent turns must be included in your token totals. No API key or model call is needed.
What changes on December 1?
| Period | Input | Output |
|---|---|---|
| Through November 30, 2026 | $2 | $10 |
| From December 1, 2026 (announced) | $4 | $20 |
At these rates, the same uncached input and output workload doubles in price. That does not mean every complete application bill doubles: cached input, infrastructure and other services have their own billing rules. Recheck Google's pricing before committing a budget.
How to reproduce the estimate
Introductory cost = input millions × 2 + output millions × 10. Later cost = input millions × 4 + output millions × 20. For example, 100,000 input tokens and 25,000 output tokens cost $0.45 initially and $0.90 at the later rates.
For a monthly forecast, record usage from representative successful tasks as well as failed attempts. Multiply by expected task count, then include retries and additional turns. A conversation can reuse context across several calls, so document length alone is not an estimate of total billed input.
Thinking and long answers affect the budget
Argon supports adjustable thinking levels. The output allowance covers thinking and output together. Use the billable usage reported by the provider rather than counting only the words visible in a final answer. Compare quality and consumed tokens on your own tasks before choosing a thinking level.
This page isolates basic token pricing. For a broader provider comparison, see the LLM API pricing hub. Subscription quotas, API spending and application hosting are separate budgeting questions.
Frequently asked questions
Does the calculator include cached input?
No. It deliberately estimates uncached input and billable output only. Do not enter cached tokens as uncached input if you need an invoice-level estimate.
Is Gemini 4 Argon available through the API?
Google's release notes list general availability on October 13, 2026, with model ID gemini-4-argon. Access and account requirements should be checked with Google.
Will this calculate my exact invoice?
No. It is a planning estimate with two published rate schedules. Actual bills depend on usage categories, service tiers and additional charges.
Sources: Google's Argon model and introductory pricing guide, official API pricing, release notes. Google controls availability and future pricing; this site is independent.