Google released its flagship Gemini 4 Argon to developers and businesses worldwide on October 13, a launch aimed squarely at winning back ground from OpenAI and Anthropic after months of delay. Google positions Argon for deep work: complex, multi-step tasks across coding, finance, law, cybersecurity and science. That framing matters. The frontier contest is no longer about a single clever answer in a chat window. It is about whether a model can stay coherent through a long job and produce work a professional can actually use.
The commercial terms will get builders' attention first. Introductory API pricing starts at 2 dollars per million input tokens and 10 dollars per million output tokens, with rates set to double after November 30. Paying customers can reach Argon through Google's AI application-building tools, while AI Pro and Ultra subscribers get it in the Gemini app and AI Mode in Search. The Antigravity coding tool is limited to Ultra subscribers. A launch price with an expiration date is a deliberate evaluation deadline: test now, while the economics are favorable, but model any production budget at the higher rate that follows.
Google's evidence so far is its own. The company says teams using Argon in Antigravity increased agentic code submissions by more than 35 percent while cutting code rollbacks by nearly a third. Early customers testing the model include Deloitte, LexisNexis, Ambience Healthcare and Isomorphic Labs, a spread that matches the pitch across professional services, legal research, healthcare and science. Treat vendor numbers as a hypothesis, not a verdict. The useful test runs on your own repository, your own documents and your own failure modes, where rollbacks and review time are costs you already measure.
Two side moves deserve attention. Google will open its Fairwind Program to hundreds of additional partners, giving vetted cybersecurity defenders a version of Argon without its cyber guardrails, an explicit bet that defenders need the same capability attackers might seek. Separately, 10,000 scientists will get six months of free access to Antigravity's highest usage tier. All of this follows a rocky stretch: Google said in September that it no longer planned to release Gemini 3.5 Pro, once expected in June, and it overhauled DeepMind after co-founder and CEO Demis Hassabis stepped aside and several Gemini leaders left.
For people building with AI, Argon changes the shortlist, not the method. Put it through a paid pilot on one real workflow, compare total cost per accepted result against your current model, and count human correction time as part of that cost. If it wins on deep, multi-step work before November 30, negotiate and plan as if the doubled price is the real one, because it soon will be. Google has finally shipped the flagship it kept promising. Now it has to prove, in production, that deep work is where it leads.
