Google Unveils Gemini 4 Argon Frontier AI Model

Google announced Gemini 4 Argon, its first new frontier AI model in nearly a year, designed for complex reasoning tasks in software engineering, finance, legal work, and cybersecurity. The model features a record 1 million output tokens and achieves state-of-the-art performance on several benchmarks. However, access is currently restricted to vetted cybersecurity experts through Google's Fairwind Program, with broader availability planned for paying API customers and Google AI Ultra subscribers.

Google unveiled Gemini 4 Argon on September 30, 2026, marking a significant return to frontier model development after a prolonged gap. The new model represents Google's answer to competitors OpenAI and Anthropic, arriving nearly a year after the previous flagship Gemini 3 generation.

Argon is engineered for sustained deep reasoning across complex, long-horizon workflows. The company claims the model delivers frontier performance across real-world software engineering, enterprise knowledge work such as legal and finance, and cybersecurity defense. Notably, Google expanded the model's output token limit to 1 million tokens-significantly higher than competitors' offerings-enabling more extensive reasoning in a single pass.

Performance benchmarks show strong results. On DeepSWE v1.1, a long-horizon software engineering benchmark, Argon scores 77.9%. On AutomationBench, Zapier's benchmark for end-to-end business execution, it ranks first at 51.3%. The model also achieves 91.7% on LVBench for long video understanding. Google says Argon outperforms OpenAI's GPT-6 Astra and Anthropic's Claude Opus 5.5 on several coding and knowledge work benchmarks.

Yet Google's rollout strategy reflects caution on safety grounds. Rather than a public release, Argon is initially available only to vetted cyber defenders participating in Google's Fairwind Program, as part of the U.S. government's voluntary pre-release access process. Google said it will expand access "as soon as possible" to paid API customers and Google AI Ultra subscribers, though no specific timeline was provided. The introductory pricing is $2 per million input tokens and $10 per million output tokens, with cached input discounted by 95%.

Google emphasized that the phased rollout reflects the company's approach to safely releasing frontier capabilities. The model is designed to autonomously find, validate, and patch critical software vulnerabilities, and includes mitigations against prompt injection attacks and misalignment. Early demonstrations showed Argon detecting critical vulnerabilities in healthcare software used globally.