Google Launches Gemini 4 Argon Frontier AI Model

Google has announced Gemini 4 Argon, a new frontier AI model initially rolling out to trusted cybersecurity defenders through its Fairwind Program. The model features an industry-leading 1 million output token limit and leads benchmarks in software engineering, finance, legal, and long video understanding. Internally, Argon has already freed over 300 TiB of memory across Google data centers and is helping migrate large C/C++ codebases to Rust. Pricing starts at $2 per million input tokens and $10 per million output tokens. Broader access for developers, enterprises, and consumers is planned following a phased release.