Google announced Gemini 4 Argon on September 30, 2026, positioning it as the lead frontier model in the Gemini 4 family with industry-leading performance on long-horizon software engineering (77.9% on DeepSWE v1.1), enterprise tasks, and cybersecurity benchmarks, outpacing recent releases from OpenAI’s GPT-6 Astra and Anthropic’s Claude Opus 5.5 in most self-reported tests. The 1M-token output limit and strong results in coding, legal/finance workflows, and vulnerability remediation drove initial trader focus on release timing. Rollout began immediately but remains gated to vetted participants in the Fairwind Program—over 650 organizations so far—while Google completes U.S. government voluntary pre-release safety reviews and iterates on guardrails. Broader developer, enterprise, and consumer access is promised “as soon as possible” in phases, with introductory API pricing at $2/$10 per million tokens. The recent cancellation of the previously promised Gemini 3.5 Pro and ongoing safety considerations around dual-use capabilities create the main near-term uncertainty for full public availability.