Google Launches Gemini 4 Argon with Up to 1 Million Output Tokens

Summary

Google has launched Gemini 4 Argon for developers, enterprise customers, and subscribers, describing it as its most capable large model for difficult coding, finance, legal, cybersecurity, and frontier-science tasks. The article says it ranks first on the Vals Index across finance, code, legal, and tax benchmarks, although no benchmark details are provided. Its main specification change is support for up to 1,048,576 output tokens, with input, reasoning, and final output sharing the same one-million-token context window. Google says Chrome engineers used the model to restructure tens of millions of lines of code, reducing work that previously took months to weeks; another overnight optimization reportedly cut memory-session preparation time by 40% and improved cold-import speed by 21%. Google is also offering 10,000 scientists six months of top-tier Google Antigravity access and expanding the Fairwind cybersecurity partnership to help find and fix critical vulnerabilities. The model adds multilayer defenses against jailbreaks and prompt injection, while its harmful-request violation rate is said to be nearly half that of the previous Gemini generation. Promotional pricing through November 30, 2026, is $2 per million input tokens, $10 per million output tokens, and $0.10 per million cached input tokens; regular prices then become $4, $20, and $0.20. Access is available through Google AI Studio, the Gemini API, Gemini applications, Search AI Mode, and Google Antigravity, depending on the user’s account or subscription.