Google releases Gemini 4 Argon to cyber defenders first

Google announced Gemini 4 Argon on 30 September 2026 as its frontier model for real-world coding, enterprise knowledge work and cyber defence. Koray Kavukcuoglu, chief AI architect at Google DeepMind, signed the announcement.

The model is not generally available. Access starts with a set of trusted cyber defenders in Google’s Fairwind Program. Google says it will extend next to paid API customers and Google AI Ultra subscribers, then to developers, enterprises and consumers in phases. For defensive security work, vetted defenders and Google’s internal teams receive Argon without cyber guardrails.

Google reports 77.9% on DeepSWE v1.1, a software engineering benchmark, 51.3% on AutomationBench, 68% on CWE-bench v1 for vulnerability remediation and 91.7% on LVBench for long video understanding. 9to5Google set the DeepSWE score against 74.2% for Claude Opus 5.5 and 74.1% for GPT-6 Astra. The output limit rises to 1 million tokens, from 64,000. As an example of long-horizon work, Google cites a migration of more than 800,000 lines of C and C++ code to Rust.

Introductory pricing is $2 per million input tokens and $10 per million output tokens.