Access status as of October 2, 2026 (checked 09:07 KST / October 2, 00:07 UTC against Google’s own announcement). We have not used or tested the model; everything below is what Google states, plus our own arithmetic where marked.
Short answer
Google announced Gemini 4 Argon on September 30, 2026. Right now it is rolling out to a set of trusted cyber defenders through Google’s Fairwind Program, plus early testers. Google says the wider release to developers, enterprises and consumers will start with paid API customers and Google AI Ultra subscribers — and it gives no date for that ("as soon as possible", "rolling out soon").
So if you saw the announcement and expected the model in your account today, that is most likely not the case yet.
Who can use it, stage by stage
| Stage | Who | Status in Google’s announcement | Date given |
|---|---|---|---|
| Announcement | public | done | September 30, 2026 |
| Early access | trusted cyber defenders (Fairwind Program); early / trusted testers | defenders: rolling out; testers: giving feedback | defenders: from the announcement; testers: not stated |
| First wider release | paid API customers and Google AI Ultra subscribers | "rolling out soon" | none |
| Broad availability | developers, enterprises, consumers | gradual expansion after testing | none |
Google describes this as a phased approach: it is taking part in the U.S. government’s voluntary process for pre-release model access and gathering feedback from early testers before widening access.
API pricing, as announced
| Introductory price | After the introductory period | |
|---|---|---|
| Input | $2 per 1M tokens | $4 per 1M tokens |
| Output | $10 per 1M tokens | $20 per 1M tokens |
| Cached input | 95% off the input price | not restated |
Google does not say how long the introductory period lasts.
Worked example (our arithmetic, not a Google quote): a month with 10M input tokens and 2M output tokens, no caching, would cost 10 × $2 + 2 × $10 = $40 at the introductory price and 10 × $4 + 2 × $20 = $80 afterwards. Check your own token mix before you budget: both announced prices double after the introductory period, so an uncached bill doubles whatever the mix, and an output token costs five times an input token at both prices.
What to do with this
- Free or other non-Ultra plans: Google’s announcement gives no plan-specific access and no date for you; it says the wider release starts with paid API customers and Google AI Ultra subscribers.
- Google AI Ultra subscribers and paid API customers: you are in the first wider group, but there is no date. Check Google’s official pages on the day you plan to switch rather than screenshots or social posts.
- Budgeting an API migration: plan with both price columns. The introductory rate is temporary and its end date is not published.
What we did not do
We did not get access to Argon, run it, or benchmark it. Capability and performance statements belong to Google and are not repeated here as findings. Community reactions were read only as a signal that people are asking about access; they are not usage data.
Source: Google, "Gemini 4 Argon: our next era of frontier intelligence" (page title "Introducing Gemini 4 Argon") — https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/ (published September 30, 2026; read October 2, 2026, 02:43 KST).
Re-checked against Google’s announcement and Google’s Gemini API pricing, models and plans pages on October 2, 2026, 09:07 KST.