Google unveils Gemini 4 Argon for cyber defenders

Google's Gemini 4 Argon builds on prior safety work with Google's Frontier Safety Framework and 3.8 Flash Cyber, and rolls out first to a narrow set of trusted cyber defenders through its Fairwind Program rather than the general public. The model is already powering internal Google workflows, including quantum algorithmic optimization that beat a published baseline by 40%, and memory efficiency agents that freed over 300 TiB with an estimated 500 TiB to 1 PiB in total savings. On the coding side, Argon agents are migrating C/C++ codebases to Rust, from core libraries like re2 and libgav1 up to the 800K+ line Fuchsia OS Zircon kernel, with rigorous auditing and emulation testing before production rollout. For libgav1, Argon replaced 32K lines of SIMD code, producing a memory-safe video decoder that runs 2.7x faster than the Rust port with identical video output. These internal examples suggest Google sees Argon less as a chat product and more as a workhorse for long, multi-step engineering tasks. The cybersecurity emphasis is notable: Argon is being released without cyber guardrails to trusted defenders and Google's own teams, and Wiz's Scan for Good used it to uncover a critical vulnerability in healthcare software that previous frontier models had missed. Whether the model's reported edge holds up beyond Google's internal tests and early partners will depend on the feedback from that initial cohort of cyber defenders and trusted testers, as well as the U.S. government's voluntary pre-release access process. Broad availability to developers, enterprises, and consumers is described as coming as soon as possible, starting with paid API customers and Google AI Ultra subscribers. For enterprise buyers, the practical question is whether Argon's benchmark leadership translates into reliable savings on code migration, vulnerability discovery, and long-horizon knowledge work, given the model is priced at an introductory $2 per million input tokens and $10 per million output tokens.