Google has launched Gemini 4 Argon, its latest reasoning model, releasing it globally on October 13 through the Gemini app, Google AI Studio, the Gemini API, and Google Cloud. The model is aimed at developers, businesses, and Pro and Ultra subscribers, and Google says it handles coding, legal, financial, cybersecurity, and scientific workflows.

[b][size=150]Availability Across Apps, APIs, and Cloud[/size][/b]

Per Google's October 13 announcement, Gemini 4 Argon is available to Pro and Ultra subscribers through the Gemini app and Search AI Mode. Developers can reach the model via Google AI Studio and the Gemini API, while businesses deploy it through Google Cloud. Google AI Ultra customers can also run Argon inside Antigravity, the company's agent-based software development environment.

Google priced the model at $2 per million input tokens** and **$10 per million output tokens. According to the developer documentation, both rates double after November 30, once the introductory period closes.
[attachment=0]gemini1.jpg[/attachment]
Specifications list a context window of 1,048,576 tokens and several levels of reasoning effort. Argon produces text output, but Google says it can take in text, images, video, audio, and PDF documents.

[b][size=150]Internal Testing Shows Coding and Reasoning Gains[/size][/b]

Google reported over 35% more agentic code submissions through Antigravity during internal testing. The company also said engineering teams reduced code rollbacks by nearly a third with Argon, and cited a 40% reduction in session preparation time for its Agent Development Kit.

Google said Argon claimed the top spot on the Vals Index, which benchmarks AI performance for professionals. Finance, legal work, coding, and tax work ranked among its strongest areas. The release includes a model card covering safety testing, evaluation methodology, and limitations.

Gemini 4 Argon arrives five days after Anthropic launched Claude Haiku 5.5, a smaller model for high-volume tasks such as customer support, priced at $0.10 per million input tokens for prompts under 100,000 tokens.

[b][size=150]Security Partners and Scientific Research[/size][/b]

In cybersecurity, Google says Argon helped researchers identify and fix software defects in a controlled environment. Through Google's Fairwind Program, Wiz probed vulnerabilities in healthcare and government systems, and the program expands to hundreds more vetted security partners starting October 13.

In scientific research, Google announced that Argon accelerated molecular modeling calculations and supported quantum computing experiments. Researchers at Isomorphic Labs used the model to review patents and scientific documents for drug discovery.

Google said safety assessments addressed biological misuse, cyber threats, and unauthorized use of autonomous agents. Testing reduced successful prompt-injection attacks to 0.7% across 15 adaptive attempts. Google also opened applications for 10,000 scientists to receive six months of free Antigravity access.

The launch follows Anthropic's $150 million commitment to federal AI research, announced October 8 as part of its role in the US Genesis Mission.

[b][size=150]Finding the Right Google Gemini Subscription[/size][/b]

As Gemini 4 Argon rolls out to paid tiers, more users are exploring plan options to access the newest reasoning features. Marketplaces like z2u.com make it simple to compare a [b][url=https://www.z2u.com/gemini/subscriptions-12-28045]Google Gemini Subscription[/url][/b] and secure a [b][url=https://www.z2u.com/gemini/subscriptions-12-28045]Gemini Premium Subscription[/url][/b] at competitive rates. Shoppers hunting for [b][url=https://www.z2u.com/gemini/subscriptions-12-28045]Gemini Subscription Deals[/url][/b] can browse[b] [url=https://www.z2u.com/]z2u.com[/url] [/b]to find discounted plans and unlock Pro and Ultra capabilities without paying full retail prices.