Google offers Gemini 4 Argon first to trusted defenders

Google is limiting Gemini 4 Argon to Fairwind cyber defenders, with a million-token figure, $2 and $10 intro pricing, and disputed benchmark leads.

Google announced Gemini 4 Argon, a frontier model for coding, enterprise work, and cybersecurity, first for trusted defenders in the Fairwind Program rather than the public. Security Affairs and CSO Online report an output-limit rise from 64,000 to 1 million tokens and introductory pricing of $2 per million input tokens and $10 per million output tokens; CSO adds that prices later rise to $4 and $20 and that Google cited a US voluntary early-access safety review. Google's September 2026 recap, published 2 October, instead describes a 1-million-token context window for coding, cyber defense, and autonomous patching, and also lists Fairwind-limited Gemini 3.8 Flash Cyber plus Gemini 3.8 Flash, Live voice models, expressive TTS, Lyria 3.5, and Connected Apps. Reported scores include 77.9% on DeepSWE v1.1, 91.7% on LVBench, 51.3% on Zapier's AutomationBench, and 68.9% on the Vals Index, with DeepSWE and Vals said to lead Claude Opus 5.5. Sources disagree on CWE-bench v1 at 68%: one Security Affairs account calls it a lead, while later wording and CSO describe a tie with Grok 4.7, GPT-6 Astra, and Claude Opus 5.5, and CSO says Claude Opus 5.5 still leads PostTrainBench at 49.3% to Argon's 45.3%. For approved security use, usual cyber restrictions are relaxed; Wiz reportedly found a critical healthcare-software flaw earlier models missed, while Google cites stronger prompt-injection resistance and real-time chain-of-thought monitoring that can stop execution.