Google DeepMind Unveils Gemini 4 Argon, Its New Frontier Model for Cybersecurity and Enterprise Work
Google's DeepMind division has unveiled Gemini 4 Argon, its new frontier model designed for long-horizon work across various industries, including software engineering, enterprise knowledge work, and cybersecurity defense. The model is not yet generally available but will be rolled out in phases to trusted cyber defenders through the Fairwind Program, which includes over 650 partners.
The introductory pricing for Argon is $2 per million input tokens and $10 per million output tokens, with these prices set to rise to $4 and $20 respectively after an initial period. The output token limit will also increase from 64K to 1M tokens during this time.
According to Google's own benchmarks, Argon outperforms OpenAI's GPT-6 Astra and Anthropic's Claude Fable 5.1 and Claude Opus 5.5 in most tests, achieving scores such as 77.9% on DeepSWE v1.1 and 91.7% on LVBench.
Argon has already demonstrated its capabilities within Google itself, freeing up over 300 TiB of data-centre memory and beating a published quantum-subroutine baseline by 40%. Additionally, it has been used in the Wiz Scan for Good initiative to identify a critical flaw exposing personal data in healthcare software.
Google is taking part in the US government's voluntary pre-release model access process and strengthening safeguards against misuse before a wider release.