
Gemini 4 Argon Debuts With Cyber Defenders First in Line
Gemini 4 Argon is Google's most powerful model yet. See the $2/$10 intro pricing, the 1M-token output limit and why cyber defenders get access first.
Google announced Gemini 4 Argon on September 30, 2026, and the company calls it its most powerful model so far. What makes the launch unusual is the order of access: the first people to get Argon are trusted cyber defenders, through a new Fairwind Program, before developers, enterprises and consumers see it.
- Argon reaches trusted cyber defenders first through the Fairwind Program; broader availability is promised "as soon as possible," with no date given.
- Google's introductory pricing is $2 per million input tokens and $10 per million output tokens, with cached input at a 95% discount.
- Standard pricing after the introductory period is $4 input and $20 output per million tokens.
- The output limit rises to 1 million tokens, up from 64,000.
What is Gemini 4 Argon?
Argon is the newest flagship in the Gemini family, aimed at software engineering, knowledge work such as legal and finance tasks, and cyber defense. Google describes the security use as autonomously finding, validating and patching vulnerabilities, which is why defenders are first in line. It builds on a busy few months for the family, including the Gemini 3.8 Live avatar release.
How does Argon score on benchmarks?
These numbers come from Google's own announcement, so treat them as the vendor's claims rather than independent results:
- DeepSWE v1.1: 77.9%
- CWE-bench v1: 68%, which Google lists as tied for first
- AutomationBench: 51.3%, listed as first
- LVBench: 91.7%
Google also says Argon leads on the Vals Index and on Gray Swan's indirect prompt-injection robustness test. According to TechCrunch, Google says Argon beat OpenAI's GPT-6 Astra and Anthropic's models on various benchmarks. Pricing comparisons are easier to check than benchmark claims; our GPT-6.1 Sol vs Astra pricing breakdown and Claude Sonnet 5.5 vs Opus 5.5 comparison give useful reference points.
What does the 1M-token output limit change?
A 1 million token output ceiling means a single response can carry an entire generated codebase section, a long legal analysis or a full remediation report without being cut off. For agent workflows that previously had to stitch many 64,000-token responses together, that is a real simplification.
Why release to cyber defenders first?
A model that can find and patch software flaws is most useful in the hands of the people who maintain that software. Starting with vetted defenders gives them a head start on hardening code before access widens. Google has not said when developers, enterprises and consumers will get Argon, so watch the AI section for the rollout. TechCrunch also reports the Gemini app has passed one billion monthly users, a figure we attribute to that outlet.
Sources: Google Blog: Gemini 4 Argon — September 30, 2026; TechCrunch — September 30, 2026.
More AI Stories

Strands Decider 2B: What Amazon's Free Decision Model Does
Strands Decider 2B is a free 2B-parameter decision model that picks options in about 115ms on a single GPU. Here is how it works and where it fits.

Meta Muse Gadgets SDK: Build Your Own AI Agent Hardware
Meta's open-source Muse Gadgets SDKs bring its Muse agent to Raspberry Pi and ESP32 builds, and 5,000 free Home Link dongles are going to subscribers.

AstaBrief 8B: Ai2's Open Model for Cited Science Reports
AstaBrief 8B is Ai2's Apache 2.0 open model that writes cited research reports in 51 seconds, 3.5x faster than its Claude pipeline. Here is how it works.
