Gemini 4 Argon launched with an introductory price that ends on November 30, 2026. Here is what it costs per token now and from December 1, what everyday jobs actually add up to, and a calculator to price your own.
Gemini 4 Argon price per million tokens
| Per 1M tokens | Until Nov 30, 2026 | From Dec 1, 2026 |
|---|---|---|
| Input (what you send) | $2.00 | $4.00 |
| Output (what Argon writes, including thinking) | $10.00 | $20.00 |
| Cached input (95% off) | $0.10 | $0.20 |
| Batch or Flex input / output | $1.00 / $5.00 | $2.00 / $10.00 |
| Priority input / output | $3.60 / $18.00 | $7.20 / $36.00 |
A token is roughly three quarters of an English word, so one million tokens is about 750,000 words. Prices don't change for very long prompts. Thinking tokens are billed as output, and caching happens automatically when you resend the same opening content.
The one number to remember: output costs five times as much as input. With long decoding, a single maximum-length Argon answer costs about $10 now and $20 from December 1. Cap your outputs.
Gemini 4 Argon cost calculator
What real jobs cost with Gemini 4 Argon
Rough examples at Google's published prices, without caching:
| Job | Tokens in / out | Intro | Standard |
|---|---|---|---|
| A normal chat answer | 2K / 500 | $0.009 | $0.018 |
| Review a 200-page contract | 100K / 5K | $0.25 | $0.50 |
| Plan a large code migration | 500K / 50K | $1.50 | $3.00 |
| Write the whole migration in one go | 500K / 200K | $3.00 | $6.00 |
| One maximum-length answer | 50K / 1M | $10.10 | $20.20 |
The pattern is clear: reading is cheap, writing is where the money goes. For most teams, a monthly Argon bill will be driven by how much it writes, not how much you feed it.
How to spend less on Gemini 4 Argon
- Cache what repeats. If you send the same codebase, manual or policy again and again, keep it at the start of the prompt: cached input is 95% off, and Google applies it automatically.
- Batch what can wait. Batch and Flex requests cost half the standard price, which suits overnight reports and bulk reviews.
- Cap the output. Set a maximum output length on every call. Raise it only for the jobs that truly need a giant answer.
- Plan, then write. A short planning call is cheap. It stops you paying for a huge answer that heads the wrong way.
- Use a smaller model for small jobs. Quick classification, short replies and simple edits rarely need a frontier model.
Is Gemini 4 Argon free?
No. Google's pricing page says Gemini 4 Argon is not available on the Gemini API's free tier, and in the Gemini app it needs a Google AI Pro or Ultra subscription. We'll update this page if a free option appears. See the release date page for where you can use it.
What about Gemini 4 Carbon's price?
Google's newer model, Gemini 4 Carbon, announced on October 28, has exactly the same intro and standard prices as Argon. It is limited to cyber defenders for now. Our sister site has the Gemini 4 Carbon pricing guide.
Making images or video? That's not billed in tokens.
Argon writes text. For pictures and video, Lumeta gives you Veo 3.1, Gemini Omni, Nano Banana 2 and more on one set of credits.
Pricing questions
How much does Gemini 4 Argon cost?
$2 per million input tokens and $10 per million output tokens until November 30, 2026, with cached input at $0.10. From December 1 it is $4 input, $20 output and $0.20 cached.
When does the Gemini 4 Argon intro price end?
November 30, 2026. The standard price starts on December 1, 2026.
How much does a 1M-token Gemini 4 Argon answer cost?
About $10 for the output at the intro price, or $20 from December 1, plus the cost of whatever you sent in.
Is there a Gemini 4 Argon free tier?
No. Argon is not on the Gemini API free tier, and in the Gemini app it needs a Google AI Pro or Ultra plan.
Sources: Gemini API pricing, Google's Argon availability post, 9to5Google. Example costs are our own arithmetic from Google's published prices.