Gemini 4 Argon API Pricing Calculator

Estimate monthly token costs for Standard, Batch, Flex and Priority, including cached input, thinking tokens and the December 1, 2026 price change. By R.K., Creator and Business Economics Analyst. Updated October 14, 2026.

Gemini 4 Argon input and output token price by tier

USD per million tokens, from Google’s Gemini API pricing page. Output includes thinking tokens. Caching is implicit only.

Which tier to use

• Standard: the normal synchronous tier, with responses in seconds to minutes.
• Batch: asynchronous bulk jobs that return within 24 hours.
• Flex: half price for synchronous calls that can wait. Google targets 1 to 15 minutes, with variable latency and best-effort availability.
• Priority: low latency and high reliability for workloads that cannot be shed, at 1.8 times the Standard rate.

How to use the calculator

1. Pick your tier. Use Standard for normal calls. Choose Batch or Flex if the work can wait, since both cost half as much.
2. Estimate monthly tokens. Multiply requests per month by tokens per request, then divide by 1,000,000 to get millions of tokens. If you have no usage data yet, run a small sample and count.
3. Split input into uncached and cached. Use the cached input token count reported in your Gemini API usage data. Do not assume every repeated part of a prompt is cached: Argon uses implicit caching, and cache hits are not guaranteed. If your usage report gives total prompt tokens and cached input tokens, subtract cached input from the total to calculate uncached input.
4. Add output including thinking. Count visible answer tokens plus thinking tokens, because Google bills both as output.
5. Read the two results. The first box is your monthly cost through November 30, 2026. The second is the cost from December 1, with the difference below it.

Example: a support assistant

A support assistant handles 20,000 requests a month. Each request sends 3,000 input tokens, and this example assumes 2,000 of them are billed as cached input. Each request returns 600 output tokens including thinking.

On Standard, that is 20 x $2 + 40 x $0.10 + 12 x $10 = $164 a month through November 30. From December 1 it is 20 x $4 + 40 x $0.20 + 12 x $20 = $328. On Batch or Flex, the same usage costs $82, then $164. Select Support bot in the calculator below to load these numbers.

Gemini 4 Argon API cost calculator

Worked monthly cost example

Take 50 million uncached input tokens and 10 million output tokens a month, with no cached input.

What changes on December 1, 2026

Every token rate on every tier doubles: Standard input goes from $2 to $4, output from $10 to $20 and cached input from $0.10 to $0.20. Your bill doubles at unchanged usage and tier. Google’s currently published pricing table does not specify a scheduled December 1 increase for grounding fees. Check the pricing page again near that date; grounding charges are separate from the token-rate calculations shown here.

How Search and Maps grounding affect cost

Grounding charges are separate from token costs and are not included in this calculator. Google’s current Gemini 4 Argon pricing table lists Search and Maps grounding for the Standard and Priority tiers. Search grounding includes 5,000 free queries per month, shared across Gemini 3 and newer models, then costs $14 per 1,000 Search queries. One API request can trigger multiple Search queries, each of which is billed. Maps grounding has a 5,000-request monthly free allowance, also shared across Gemini 3 and newer models, then costs $14 per 1,000 Maps search queries. Check Google’s pricing page before budgeting, as these rates can change.

Release date, access and model limits

Sources: Google’s Gemini 4 Argon model page, Gemini API pricing and the DeepMind model card.

Argon price compared with other models

List prices in USD per million tokens, October 2026. This compares token prices, not model quality or cost for equivalent work. Tokenization, context length, caching, output volume and tool use change the final bill. OpenAI charges more for long-context requests, so check its pricing page.

Cheap per token is not cheap per task

Argon and GPT-6.1 Sol list at the same token prices, but that does not make a task cost the same. Google bills thinking tokens as output, and reasoning models spend different numbers of tokens on the same job. Run a sample of your own work and count tokens before you compare models. Any published cost-per-task figure reflects one benchmark workload and one set of settings, so use it as a prompt to measure, not as a verdict.

What to do before December 1

• Run a real sample of your workload now and record uncached input, cached input and output tokens, including thinking.
• Plug those numbers into the calculator and budget at the December 1 column, not the introductory one.
• Use Batch or Flex for work that can wait. Both are half the Standard rate, so Argon costs $1 input and $5 output until November 30.
• Test a cheaper model on the easy steps. Gemini 3.8 Flash lists at $0.75 input and $3.75 output through December 31, 2026, then doubles on January 1, 2027.
• Put the long, repeated part of each prompt first. Argon caching is implicit, so you cannot force it.

Frequently asked questions

When was Gemini 4 Argon released?

Google announced Gemini 4 Argon on September 30, 2026, with access limited to approved cybersecurity defenders. It became generally available in the Gemini API on October 13, 2026 under the model ID gemini-4-argon.

How much does the Gemini 4 Argon API cost?

Standard pricing is $2 per million input tokens, $0.10 per million cached input tokens and $10 per million output tokens through November 30, 2026. From December 1 it is $4, $0.20 and $20. Batch and Flex cost half the Standard rates. Priority costs $3.60, $0.18 and $18 through November 30, then $7.20, $0.36 and $36.

Does the December 1 increase apply to every tier?

Yes. Token rates double on Standard, Batch, Flex and Priority, including cached input. A workload that costs $200 a month on Standard at introductory rates costs $400 from December 1 if usage stays the same.

Is Gemini 4 Argon free?

No. Google’s API pricing page says Argon is not available on the Free tier or on the Google AI Pro plan. Consumer Gemini app subscriptions are billed separately from API usage.

Are thinking tokens billed?

Yes. Google bills thinking tokens as output. Argon supports low, medium and high thinking levels, and higher levels can raise your output token count, so estimate output from real runs.

What is the difference between Batch and Flex?

Both cost half the Standard rate. Batch is asynchronous, with results in up to 24 hours. Flex is synchronous with a target of 1 to 15 minutes, variable latency and best-effort availability, so requests can fail when capacity is tight.

Do Google Search and Maps grounding cost extra?

Yes, and they are separate from token charges. Google lists 5,000 free Search grounding queries per month, shared across Gemini 3 and newer models, then $14 per 1,000 queries. One request can trigger several queries, and each is billed. Maps grounding has its own 5,000 free requests and the same $14 per 1,000 rate after that.

Is Gemini 4 Argon the same as Gemini 4 Pro?

As of October 14, 2026, Google’s Gemini API documentation lists Gemini 4 Argon (gemini-4-argon). The current API documentation does not list a separate Gemini 4 Pro model. Check Google’s current model list rather than assuming that Gemini 4 Pro is an alias for Argon.

Methodology: Pricing checked against Google’s Gemini API pricing page and the Gemini 4 Argon model page on October 14, 2026. Estimates use published per-million-token list rates for Standard, Batch, Flex and Priority. Search and Maps grounding, taxes and other charges are excluded. Prices can change, so verify before you commit.