Google Prices Gemini 4 Argon at $2. Alphabet Closed at…
Google set introductory pricing for Gemini 4 Argon at $2 per million input tokens and $10 per million output tokens, but access is initially limited to trusted cyber defenders in its Fairwind Program.
On September 30, 2026, Google announced Gemini 4 Argon, a frontier AI model that will initially go to a limited group of trusted cyber defenders through its Fairwind Program. Koray Kavukcuoglu, senior vice president at Google DeepMind and the company's chief AI architect, said developers, enterprises and consumers would get access as soon as possible, after additional guardrails. The model is not a public product launch, and the rate card is public while the meter is not.
Pricing Without a Public Meter
Google published an introductory rate card of $2 per million input tokens and $10 per million output tokens, with cached input at 95% off. After the introductory period, the listed price doubles to $4 per million input tokens and $20 per million output tokens. Google also said the output-token limit rises from 64,000 to 1 million, a 15.625x increase. The blog post did not describe that change as a context-window specification.
Google claimed top scores on several benchmarks, including 77.9% on DeepSWE v1.1 and a tie for first at 68% on CWE-bench v1. Those are company figures, not independent audits. Wiz, through Scan for Good, was named as an early user. Tulsee Doshi, product lead for Gemini, told CNBC the limited rollout gives Google more confidence while putting a strong cyber-defense model in defenders' hands quickly.
Alphabet's stock gave traders a number. Nasdaq's regular-session close for GOOGL was $344.08 on September 30, and $343.50 on Friday, October 2, up 1.56% that day. Nasdaq listed a market capitalization of $4.201 trillion. Investing.com reported an after-hours gain of about 2% on September 30. On October 1, GOOGL opened at $350.785 and closed at $338.24.
The launch came one day after Sundar Pichai signed a voluntary White House AI safety accord alongside leaders from Anthropic, OpenAI, Meta, xAI and Nvidia. Google separately cited a U.S. voluntary pre-release process for the model. Commercial access, Google said, will start with paid API customers and Google AI Ultra subscribers rather than a free switch in the Gemini app. JPMorgan analysts said Google still needs a step forward in personal agents, while Bank of America pointed to Google Cloud and existing products. CNBC reported the introductory rates match OpenAI's discounted GPT-6.1 Sol.
Because customers outside Fairwind cannot call the model, Argon consumption cannot enter Google Cloud revenue for the current quarter. Once keys exist, cache discounts and a 1 million output-token cap could become competitive points. Google has not dated when the introductory pricing ends or when broader access begins.
Source: FinanceFeeds