The Gemini 4 Argon release date for public access is October 13, 2026. Google DeepMind announced the model on September 30 for Fairwind cyber defenders only; two weeks later it opened Argon globally for paid Gemini API customers, Google AI Ultra users in Antigravity and AI Studio, and Google AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search.
This page updates the earlier "no public date yet" status with who can use Argon today, the confirmed model ID and token limits, introductory API pricing through November 30, 2026, and how Fairwind still differs from the public build. Facts below are checked against Google's October 13 launch posts and the Gemini API docs as of October 15, 2026.
Gemini 4 Argon release date: what changed on October 13
On September 30, 2026, Google said Argon would roll out in phases, starting with trusted cyber defenders in the Fairwind Program, then paid API customers and Google AI Ultra subscribers, then a broader audience when safeguards were ready. It gave no calendar date.
On October 13, 2026, Google published that Argon is "now globally available" for those next groups. The same day, the Gemini API release notes listed gemini-4-argon as generally available (GA), and the model appears on the models and pricing pages. A matching developer post points builders to the paid API tier, AI Studio, Android Studio, and Antigravity.
So the short answer to "when is the Gemini 4 Argon release date?" is: public GA on October 13, 2026, after a two week Fairwind only window that began September 30.
Who can use Gemini 4 Argon today
Google's October 13 post breaks access into clear buckets:
| Who | Where Argon is available | Notes |
|---|---|---|
| Everyday subscribers | Gemini app and AI Mode in Search | Google AI Ultra or Pro |
| Developers (paid API) | Gemini API, Google AI Studio, Android Studio | Paid tier; not Free |
| Google AI Ultra (dev tools) | AI Studio and Google Antigravity | Highest Studio quota; Argon in Studio |
| Enterprises | Gemini for business, API, Antigravity | Same model ID |
| Scientists / researchers | Antigravity highest usage tier | Apply for 6 months free (eligibility, capacity, selection) |
| Cyber defenders | Fairwind Program | Expanding; Argon without cyber guardrails |
Important Studio caveat from Google's AI plans docs: Gemini 4 Argon is not available on the Free or Google AI Pro plans inside AI Studio. Pro subscribers still get Argon in the consumer Gemini app and AI Mode; Ultra is the consumer plan that also includes Argon in Studio and Antigravity.
Consumer plan list prices on Gemini subscriptions remain AI Plus at about $4.99, AI Pro at $19.99, and AI Ultra at $99.99 or $199.99 a month in the US depending on usage level. Always check Google's live pricing page for your region.
If you only need a free chat surface, Argon is not on Free. For a wider flagship comparison, see ChatGPT vs Claude vs Gemini vs Grok. ChatGPT and Claude still matter when you want a non-Google top model on a paid plan today.
Gemini 4 Argon API: model ID, limits, and pricing
Model ID: gemini-4-argon (stable), per the Gemini 4 Argon model page.
Token limits (official):
| Limit | Value |
|---|---|
| Unified context (input + output) | 1,048,576 tokens |
| Default max output (synchronous) | 65,536 tokens |
| Long decode max output | up to 1,048,576 tokens |
Inputs include text, image, video, audio, and PDF; output is text. Thinking levels (low, medium, high), search grounding, code execution, function calling, and computer use (preview) are supported. Image generation and Live API are not. Caching is implicit only (explicit caching is not supported).
Standard paid API price from the Gemini API pricing page:
| Period | Input / 1M tokens | Output / 1M tokens | Implicit cache / 1M tokens |
|---|---|---|---|
| Through November 30, 2026 | $2.00 | $10.00 | $0.10 |
| From December 1, 2026 | $4.00 | $20.00 | $0.20 |
Batch and Flex tiers are half the standard rates during the same windows; Priority is higher. Google's launch footnote states the introductory period ends November 30, 2026, and the higher price applies to the public Argon model only. Argon is not on the Free API tier.
Pros: GA model ID you can ship against, 1M unified context with long decode, clear intro price window, Pro/Ultra consumer access in the Gemini app. Cons: not on Free or Pro in AI Studio, output can be long and expensive, price doubles after November 30, Fairwind cyber build stays gated.
What Argon is built for (vendor claims, attributed)
Google positions Argon for long horizon software engineering, enterprise knowledge work (legal, finance, science), and cybersecurity defense. Headline claims from the September 30 and October 13 posts include:
- Up to 1M output tokens with the new long decode / continuation setup (far above the prior 64K generation limit).
- 77.9% on DeepSWE v1.1 (Google reported long horizon software engineering).
- #1 on the Vals Index and 51.3% on Zapier's AutomationBench (Google reported).
- 91.7% on LVBench long video understanding (Google reported).
- 68% on CWE-bench v1 for vulnerability remediation, tied for first (Google reported).
Independent Artificial Analysis earlier scored Argon 53 on its Intelligence Index, level with GPT-6 Astra and well above Gemini 3.1 Pro Preview, with a lower hallucination rate but much higher average output tokens per task. Treat vendor and third party scores as directional, not a single ranking of "best model." For how labs describe safety gates around frontier releases, see Preparedness Framework vs Responsible Scaling Policy. Google DeepMind publishes Argon's model card and eval methodology alongside the consumer and API rollouts.
Fairwind vs the public model
Fairwind is still the path for vetted cyber defenders who need Argon without cyber guardrails. Google's October 13 post says it is expanding Fairwind to hundreds of additional partners and that Wiz continues to use Argon in Scan for Good for critical infrastructure. Public API and consumer Argon keep misuse, jailbreak, prompt injection, and misalignment monitors in place under Google's Frontier Safety Framework.
Scientists can also apply for Antigravity's highest usage tier free for six months (rolling onboarding, subject to eligibility, capacity, and selection). That is a research access path, not the same as Fairwind's ungated cyber build.
How Argon fits next to other frontier releases
October 2026 already stacked several public frontier launches: GPT-6 Sol/Luna in ChatGPT, Claude Haiku 5.5, and Mistral Large 4 in preview. Argon's public GA closes the gap Google left open on September 30, when only Fairwind could touch the model. If you are choosing among top paid assistants rather than wiring the API, start with ChatGPT vs Claude vs Gemini vs Grok and our GPT-6 Astra vs Claude Fable 5.1 comparison for the highest tier API pair.
Quick checklist: are you on Argon yet?
- Look for model ID
gemini-4-argonin the Gemini API / AI Studio picker (paid project or Ultra Studio access). - In the Gemini app or AI Mode, confirm you are on AI Pro or AI Ultra (not Free or Plus alone).
- For cyber work without guardrails, apply to Fairwind rather than assuming the public API is ungated.
- Budget for the December 1, 2026 price step up if you are shipping long decode workloads.
From Ahzy Capital (our own products, labeled as such): Get the free AI Agent Cheat Sheet: 24 copy and paste AI agent setups for 8 types of small business, with prompts and workflows. We make and sell these, so they aren't independent picks.
For a short spec sheet, see our Gemini 4 Argon model profile, and for other launches, the AI model release timeline.
FAQ
When is the Gemini 4 Argon release date?
October 13, 2026 for public GA (API paid tier, Ultra in Antigravity/AI Studio, and Ultra/Pro in the Gemini app and AI Mode). Google first announced Argon on September 30, 2026 for Fairwind cyber defenders only.
Is Gemini 4 Argon free?
No. It is not on the Free Gemini API tier or Free/Plus-only consumer access. AI Pro and Ultra subscribers can use it in the Gemini app and AI Mode; AI Studio access to Argon requires Ultra (or a paid API key), not Pro or Free.
How much does Gemini 4 Argon cost on the API?
Introductory standard rates are $2 per million input tokens and $10 per million output tokens through November 30, 2026, then $4 and $20 from December 1. Implicit context cache is $0.10 then $0.20 per million tokens. See Google's pricing page for Batch, Flex, and Priority.
What is the Fairwind Program after the public launch?
It remains Google's trusted access program for vetted cyber defenders. Fairwind users get Argon without cyber guardrails for defensive work; the public model keeps those guardrails. Google is expanding Fairwind to more partners.
What is the Gemini 4 Argon model ID?
gemini-4-argon. Default synchronous max output is 65,536 tokens; long decode can go up to 1,048,576 within a 1,048,576 unified context window.
Is Gemini 3.5 Pro cancelled?
Google has not said so. Its last clear public line (July 2026) was that 3.5 Pro was testing with partners. Gemini 3.1 Pro Preview remains listed among Gemini 3 models, while Argon is now the flagship Gemini 4 reasoning model.
