Google Begins Broad Gemini 4 Argon Rollout

Sundar Pichai hit post on X on October 13, 2026, and Gemini 4 Argon moved at once onto every paid surface he named.

“Gemini 4 Argon is here. It's available globally today for developers and enterprises in the paid tier of the Gemini API and the Gemini platform, for Ultra subscribers in Google Antigravity, and Ultra and Pro subscribers in the Gemini app and Search. Hope you all enjoy it!”

Thirteen days earlier the door had been far narrower. On September 30, 2026, the model left Google through a single channel: a set of trusted cyber defenders in the Fairwind Program, without its cyber guardrails, and nowhere else. Developers had no API. Subscribers had no app path. Google said the wider release would come as soon as possible and would begin with paid API customers and Google AI Ultra subscribers.

The October 13 blog.google post, titled “Gemini 4 Argon: Built to solve your hardest problems,” listed Pro subscribers beside Ultra, a step beyond that September promise. According to Koray Kavukcuoglu, SVP at Google DeepMind and Google’s chief AI architect, Argon excels at demanding, multi-step work in coding, finance, law, cybersecurity, and science.

What the announcement left unsaid was what the model had already done inside the company.

Google’s own software engineers had been using Argon for weeks on internal code through Antigravity. Agentic code submissions rose more than 35 percent. Code rollbacks fell by nearly a third. On the Vals Index, which scores models on finance, coding, legal, and tax work, Argon ranked first. It scored 87 percent on Embodied Reasoning, ERQA.

The Chrome team handed it a roughly six-million-line legacy folder of browser features held inside Chromium’s forty-million-line codebase. Argon helped separate core webpage handling from that folder. Piping multiple clean code changes at a time, it carried an architectural refactor across hundreds of files. The changes were built to let live webpages move fluidly between tabs, side panels, and floating windows without reloading, eliminating an entire category of browser crashes. The work finished in weeks instead of months.

That same model was now reaching the Gemini app and the paid subscriber surfaces that opened the day of Pichai’s post.

Robby Stein, vice president of product for Google Search, wrote on X that Argon “is now rolling out globally in Search.” It “brings big improvements to tackling nuanced, multi-step tasks in AI Mode,” he added, and is “better at understanding your intent and following your instructions in back and forth conversations.” Subscribers reached it through the “+” icon in AI Mode’s Ask-anything bar, the same menu that had added 3.7 Flash in August and 3.8 Flash in September.

Google’s launch post said the model “follows style instructions better than any of our previous Gemini models.” The company blog described the Gemini app arrival in one stretch: “Gemini 4 Argon is now available in the Gemini app, delivering frontier reasoning and improved performance across complex tasks. Argon answers questions more directly with less hedging, doesn't repeat itself more than necessary, and skips clichés throughout the responses. For longer tasks, it keeps you in the loop by sharing its progress more often and more effectively.”

On macOS the Gemini app also surfaced Antigravity’s agentic coding interface. A simple toggle moved users from chat into tasks the model could keep running across Google apps even after the person left the computer.

Developers paid $2 per million input tokens and $10 per million output tokens through the Gemini API, rates that held through November 30, 2026. Input tokens are the information supplied to the model; output tokens are the content it generates. Cached input tokens carried a 95 percent discount for the same stretch. After the promotional window closed, standard rates moved to $4 per million input tokens and $20 per million output tokens. The API carried a combined input and output limit of one million tokens.

Google AI Pro cost $19.99 a month and included 5TB of storage, YouTube Premium Lite, Google Health Premium, and Google Home Premium Standard. Google AI Ultra cost $99.99 a month and included 20TB of storage, YouTube Premium, Google Home Premium Advanced, and Antigravity access. The model linked to more than forty third-party apps, among them Shopify, Instacart, OpenTable, and Adobe. Tulsee Doshi, Gemini product lead at Google DeepMind, told Business Insider the model was a “powerhouse.”

Fairwind itself was growing. Google was expanding the program from the original set of trusted cyber defenders to hundreds of additional partners, who would receive Argon under the same terms: vetted access without cyber guardrails. On a parallel track the company offered ten thousand scientists six months of free access to the highest Antigravity usage tier. Applications opened; places depended on eligibility, capacity, and selection.

More models in the Gemini 4 family would follow. Launch language said Google looked forward to bringing the rest of the line.

The consumer Search surface was still catching up. Argon was meant to appear for Pro and Ultra subscribers under the “+” icon in AI Mode’s Ask-anything bar. As of 12:17 p.m. ET on day one, October 13, the model had not yet filled into that menu on at least one Pro account.