Google introduces Gemini 4 Argon for coding and cybersecurity
Google says Gemini 4 Argon targets complex coding, enterprise knowledge work and cybersecurity, with initial access limited to trusted testers through Fairwind.
Google has introduced Gemini 4 Argon for complex coding, enterprise knowledge work and cybersecurity tasks. The company says the model supports up to one million output tokens, giving it room to work through long, multi-step problems in a single run; that capacity does not by itself guarantee correct results.
Initial access is being provided to trusted testers through Google’s Fairwind Program. According to Google, a broader rollout for developers, enterprises and consumers is planned for later, so full public availability had not begun at the time of the announcement.
