In this article
Updated 13 October 2026. This is a developing story; we will update it as new official information is published.
Gemini 4 Argon is Google’s new flagship AI model, released on 13 October 2026. It is available to developers through the Gemini API and to paying Google AI Pro and Ultra subscribers in the Gemini app. Google says it is built for long, complex tasks in coding, finance, law, cybersecurity and science.
Key facts
- Released: 13 October 2026, globally.
- Who gets it: developers and businesses via the API; Google AI Pro and Ultra subscribers in the Gemini app and AI Mode in Search.
- Price: $2 per million input tokens and $10 per million output tokens, an introductory price.
- Claims: Google says Argon ranks first on the Vals Index of work tasks. That is Google’s claim and needs independent testing.
What happened
In a blog post on 13 October, Koray Kavukcuoglu, Google DeepMind’s senior vice president and Google’s chief AI architect, announced Gemini 4 Argon. Google calls it its “latest frontier model”, meaning its most capable one.
It is available in Google AI Studio and the Gemini platform for developers, and in Google’s Antigravity coding tool for Ultra subscribers. Google also published a separate version for robotics.
Who can use Gemini 4 Argon, and what it costs
| Who | Where | Cost |
|---|---|---|
| Developers and businesses | Gemini API, Google AI Studio, Gemini platform | $2 per million input tokens, $10 per million output tokens |
| Google AI Ultra subscribers | Gemini app, AI Mode in Search, Antigravity | Included in subscription |
| Google AI Pro subscribers | Gemini app, AI Mode in Search | Included in subscription |
| Free Gemini users | Not listed in Google’s announcement | n/a |
What does $2 per million tokens mean?
A token is a piece of a word; one million tokens is roughly 750,000 English words. So reading a 100-page report (about 50,000 words, or 67,000 tokens) would cost about 13 cents in input, plus the cost of the answer at $10 per million output tokens. Google calls this an introductory price, so it may rise.
What Google says it can do
- Large coding projects: Google says Argon helped its Chrome team separate a legacy folder of about 6 million lines of code inside Chromium’s 40-million-line codebase, finishing months of work in weeks.
- Engineering speed: Google says teams using Argon in Antigravity increased code submissions by more than 35% while cutting rollbacks by nearly a third.
- Work tasks: Google says Argon ranks first on the Vals Index, which tests AI on finance, coding, legal and tax work.
- Science and security: Google says researchers have used it in quantum computing, nuclear fusion and life sciences, and to find and fix security vulnerabilities.
Treat with caution: These results come from Google itself. Company benchmarks often use the settings that suit their own model best. Wait for independent tests before deciding Argon is better than rival models for your work.
Why it matters
The release continues a fast race between Google, OpenAI, Anthropic and others to build AI that can handle long, multi-step work rather than single questions. For businesses, the low introductory price is a clear attempt to win developers. For everyday users, the main change is a stronger model inside the Gemini app and Google Search for paying subscribers.
Before relying on any AI model, read our guide to what large language models can and cannot do. If you make AI videos, see our Higgsfield AI review, a platform that bundles Google’s image models.
Frequently asked questions
When was Gemini 4 Argon released?
Google released it globally on 13 October 2026.
Is Gemini 4 Argon free?
Google’s announcement lists it for Google AI Pro and Ultra subscribers and for developers through the paid API. Free users are not mentioned.
How much does Gemini 4 Argon cost?
Through the API, the introductory price is $2 per million input tokens and $10 per million output tokens.
Is Gemini 4 Argon better than other AI models?
Google says it leads several benchmarks, but these are company claims. Independent testing will give a clearer picture.
Browse more in our AI section.



