Google Returns Gemini Model After Seven-Month Drought

Google has announced Gemini 4 Argon, its first frontier model since July, aimed at software engineering, legal and finance work, and cyber defense. The new model is released in stages through Google's Fairwind Program, with vetted cyber defenders getting early access.

The initial cohort includes over 650 participating partners, who must limit access to internal security, incident-response, or penetration-testing teams and use multi-factor authentication. For this group and Google's internal teams, Argon ships without its usual cyber guardrails.

Google has raised the output limit of Gemini 4 Argon to 1 million tokens from 64,000, with introductory pricing at $2 per million input tokens and $10 per million output tokens. This will rise to $4 and $20 after the introductory period, although cached input is priced 95% below standard input rates.

Independent tests show Gemini 4 Argon is on par with OpenAI's best model but behind Anthropic's. It has reported a state-of-the-art 77.9% on DeepSWE v1.1, a long-horizon software engineering benchmark, and leading results on the Vals Index, Vals Finance Agent v2, and Harvey's Legal Agent Benchmark.

Google is also participating in the U.S. government's voluntary pre-release model access process. Broader access will follow once it has gathered feedback from early testers and strengthened its safeguards.