Gemini 4 Argon: Access, Pricing and the Dec 1 Price Rise

Illustration by OneAppleFall. Not an official Google image.

Researched October 20, 2026

Google’s newest flagship AI model, Gemini 4 Argon, became generally available on October 13, 2026, two weeks after Google first gave it to a small group of cybersecurity defenders. [1][5] It is now in the Gemini app for paying subscribers and in the Gemini API for developers.

If you are a developer or a business, the date that matters most is December 1, 2026, when Argon’s introductory API prices double. [1][3] This guide covers who can use Argon, what it costs, and what Google says it can do.

Quick answer: In the Gemini app and AI Mode in Search, Argon is for Google AI Pro and Ultra subscribers; Ultra subscribers also get it in Google Antigravity, Google’s coding tool. [1][2] For developers, it costs $2 per million input tokens and $10 per million output tokens until November 30, then $4 and $20 from December 1, 2026. It is not on the Gemini API free tier. [1][3]

This is a researched explainer based on Google’s announcements, pricing page and model card. Benchmark figures are Google’s own and have not been independently verified.

When was Gemini 4 Argon released?

Date	What happened
September 30, 2026	Google announced Argon and released it first to trusted cyber defenders in its Fairwind Program. [5][6]
October 13, 2026	General availability for developers, businesses and Pro and Ultra subscribers. [1][2]
November 30, 2026	Last day of introductory API pricing. [1][3]
December 1, 2026	API prices double. [3]

Google called the staged release necessary because “safely releasing frontier capabilities at this level requires a phased approach.” [5]

Who can use Gemini 4 Argon?

Where	Who gets it
Gemini app	Google AI Pro and Ultra subscribers [1][2]
AI Mode in Google Search	Google AI Pro and Ultra subscribers [1]
Google Antigravity (coding)	Google AI Ultra subscribers and enterprises [1][5]
Gemini API and Google AI Studio	Developers on the paid tier [1][3]
Gemini for business	Enterprises [4]
Fairwind Program	Approved cybersecurity defenders, who can get Argon without its cyber guardrails [1]

Free Gemini users do not get Argon. On the developer side, Google’s pricing page states plainly that “Gemini 4 Argon is not available on the Free tier.” [3]

Google’s announcements do not list countries or languages for the Gemini app rollout, so if you are in the UK or Europe, check the model list in your own Gemini app after subscribing rather than assuming it is there. [2]

How much does Gemini 4 Argon cost?

For subscribers, Argon is included in the Google AI Pro or Ultra plan you already pay for; Google has not announced a separate charge in the app. [1][2]

For developers using the API, prices are per million tokens. These come from Google’s official pricing page: [3]

Tier	Until Nov 30, 2026 (input / output)	From Dec 1, 2026 (input / output)
Standard	$2.00 / $10.00	$4.00 / $20.00
Batch	$1.00 / $5.00	$2.00 / $10.00
Flex	$1.00 / $5.00	$2.00 / $10.00
Priority	$3.60 / $18.00	$7.20 / $36.00

A few details that affect real bills:

Output prices include “thinking” tokens, the model’s internal reasoning, so long reasoning answers cost more than their visible length suggests. [3]
Cached input is much cheaper. Google says cached input tokens are 95% off at launch, and context caching is automatic for this model. [1][3]
Search grounding is charged separately. You get 5,000 free Google Search requests a month, shared across Gemini 3 and newer models, then $14 per 1,000. [3]

What the December price rise means

At the standard rate, a job that uses 10 million input tokens and 2 million output tokens costs about $40 before December 1 and about $80 after it. If you are planning a big batch job, an evaluation or a migration, running it in November costs half as much. You can estimate your own monthly bill, and compare Argon with other models, in our free AI API cost calculator.

For context, the post-December price of $4 / $20 matches Anthropic’s Claude Opus 5.5, while OpenAI’s GPT-6.1 Sol costs $2 / $10 and the top models from Anthropic and OpenAI cost $10 / $50, according to Yahoo Finance’s comparison. [6]

What can Gemini 4 Argon do?

According to Google’s model card, Argon accepts text, images, audio and video, answers in text, and has a combined input and output limit of up to 1 million tokens. Its knowledge cut-off is March 2026. [4] Google says its long output, up to 1 million tokens through the API, is “industry-leading.” [1][5]

Google aims it at long, multi-step work: [1][4]

software engineering and code migration;
legal and financial analysis and other “knowledge work”;
cybersecurity defence;
science and mathematical reasoning;
analysing long, mixed-media documents.

In the Gemini app, Google says Argon gives more direct answers with less hedging and repetition, and shares progress more often on long tasks. [2]

Google’s performance claims

Google says Argon ranks first on the Vals Index, which covers finance, coding, legal and tax work, and scored 87% on the ERQA embodied-reasoning benchmark, ahead of other frontier models. [1] It also describes internal projects such as helping refactor parts of Chromium. [1]

These are Google’s own figures. Bloomberg reported that some Google employees doubted Argon’s real-world coding performance compared with rivals, which Google disputes. [5][6] The fair reading is that Argon is a serious frontier model, but you should test it on your own tasks before switching.

Safety and limitations

Google calls Argon’s protections “our most comprehensive safety and security stack yet,” including layered defences against misuse, a real-time monitor that stops agents taking unauthorised shortcuts, and training against prompt injection. [1]

The model card is also candid about risks: Google’s frontier safety tests could not rule out Argon reaching its “critical capability levels” for biology and cybersecurity, so extra safeguards were applied. Like any large model, it can still produce confident mistakes (“hallucinations”) and may occasionally be slow or time out. [4]

Should you switch to Gemini 4 Argon?

Everyday Gemini users: you need Google AI Pro or Ultra. If your current plan already covers your needs, there is no rush.
Developers: try it on your own workload now, while it costs $2 / $10, and decide before December 1 whether it is worth $4 / $20 for you.
Businesses: check whether it is enabled in Gemini for business, and look at batch and flex pricing for jobs that do not need instant answers.
Cost-sensitive work: for simpler, high-volume tasks, Google’s Flash models remain much cheaper than Argon. [3]

Questions

Is Gemini 4 Argon free?

No. In the Gemini app it needs a Google AI Pro or Ultra subscription, and it is not available on the Gemini API free tier. [2][3]

When does the Gemini 4 Argon price go up?

On December 1, 2026. Standard API prices rise from $2 to $4 per million input tokens and from $10 to $20 per million output tokens. [3]

What is the Fairwind Program?

A Google programme that gives approved cybersecurity defenders early access to Argon without its cyber guardrails. Applications are accepted on a rolling basis. [1]

What is Gemini 4 Argon’s context window?

Up to 1 million tokens, counted across input and output combined. [4]

Is there a Gemini 4 Pro?

Google’s announcements name only Gemini 4 Argon so far. Treat claims about other Gemini 4 versions as rumours until Google confirms them. [1]

Gemini 4 Argon: Who Can Use It, What It Costs, and the December Price Rise

When was Gemini 4 Argon released?

Who can use Gemini 4 Argon?