< Back to all clusters
[TECHNOLOGY] · United States · 100 sources

started · updated

Google unveils Gemini 4 Argon frontier AI model

Google has unveiled Gemini 4 Argon, its most advanced frontier AI model to date, specifically engineered for complex, multi-step reasoning in software engineering, cybersecurity, finance, and law. The model features a significant technical leap, increasing the maximum output limit from 64,000 to 1 million tokens, allowing it to process and generate massive codebases and extensive professional documents.

Argon is currently being deployed in phases. It is initially available to a select group of cybersecurity professionals through the Fairwind Program and is being used internally by Google for tasks such as optimizing quantum computing algorithms and managing data center memory. Internal applications have already freed up over 300 TiB of memory. Google is also participating in a voluntary safety evaluation process with the U.S. government before a wider release.

In performance benchmarks, Argon achieved a score of 77.9% on the DeepSWE v1.1 software engineering benchmark and tied for first place on the CWE-bench v1 cybersecurity benchmark with a score of 68%. Introductory API pricing is set at $2 per million input tokens and $10 per million output tokens.

Entities

Alphabet · Cox Automotive · Donald Trump · Fairwind Program · Gemini 3.8 Live · Gemini 4 Argon · Google · Google Cloud · Google DeepMind · Koray Kavukcuoglu · Sundar Pichai · Tulsee Doshi

Claims

What the coverage asserts, and how many sources carry each claim.

Sources

about 5 hours ago
Anuncia Google Gemini 4 [massinformacion.com.mx]
about 12 hours ago
about 13 hours ago
about 20 hours ago
about 17 hours ago
about 22 hours ago
Google launches Gemini 4 Argon [www.constellationr.com]
1 day ago
about 5 hours ago
Top AI Models Every Few Days [www.nextbigfuture.com]
1 day ago
about 7 hours ago
1 day ago