AI & SITrending
Google has released Gemini 4 Argon to developers and subscribers, offering a 1 million token context window and introductory pricing of $2 per million input tokens.
Alphabet’s Google released its flagship AI model, Gemini 4 Argon, on October 13, 2026, making it available to developers and paying subscribers worldwide. The launch follows an initial restricted rollout to select cybersecurity partners and internal teams, aiming to expand access across coding, finance, law, and scientific research.
The model is available to developers through the paid tier of the Gemini API in AI Studio and Android Studio. According to Google, the Gemini app and AI Mode in Search are accessible to Google AI Pro and Ultra subscribers. Access to the Antigravity coding platform is restricted to Ultra subscribers only.
At launch, Argon carries a price of $2 per million input tokens and $10 per million output tokens. Cached input tokens are discounted by 95%. Google states that these rates are set to double after November 30, when the introductory period ends.
Google describes Argon as designed for multi-step tasks and reports a unified input-output limit of 1 million tokens via the Gemini API. The company says the model ranks first on the Vals Index, which measures performance across finance, legal, coding, and tax work.
In internal tests, Google reported that teams using Argon within Antigravity saw agentic code submissions climb more than 35%, with code rollbacks falling by nearly a third. The company also reported scores of 77.9% on DeepSWE v1.1 and 51.3% on AutomationBench. These results are company-reported and have not been independently verified.
Google is expanding its Fairwind Program to hundreds of additional cybersecurity partners. These partners receive access to a version of Argon without its standard cyber guardrails. The company says researchers have used the model to audit critical codebases and identify zero-day vulnerabilities.
Separately, Google announced it will offer 10,000 scientists six months of free access to Antigravity's highest usage tier, with applications now open. Enterprise customers testing the model early include Deloitte, LexisNexis, Ambience Healthcare, and Isomorphic Labs.
Google described Argon as its most robust safety implementation to date, stating that new jailbreak-resistance measures cut violations on an adversarial test set by close to 50% relative to earlier Gemini models. The release follows a period of delays and leadership changes at Google DeepMind, where Demis Hassabis stepped aside as chief executive. Google also confirmed it scrapped plans for Gemini 3.5 Pro, a model CEO Sundar Pichai had previously indicated would arrive as early as June.
The release of Gemini 4 Argon marks a significant expansion of Google's frontier AI capabilities into the developer and enterprise markets. By offering a 1 million token context window and specialized tools for coding and cybersecurity, Google aims to compete directly with other major AI providers. The introductory pricing strategy and free access for scientists suggest a focus on rapid adoption and ecosystem building, while the restricted access to certain safety guardrails highlights the ongoing tension between capability and security in advanced AI deployment.
Sources
This article was drafted with AI assistance and checked against the sources above. Company claims are reported as claims. Cover image is AI-generated.
The day's checked news, by email
One short email a day with the top stories. Free, and you can unsubscribe in one click.
Or follow for the latest news:
No comments yet. Start the discussion.
Microsoft introduced new security features to monitor AI agents, warning that autonomous systems can pose insider risks even without being hacked. The company urges separating intelligence supply from control authority.
The Department of Homeland Security has ordered federal civilian agencies to remove DeepSeek AI products from IT systems and barred future procurements, citing cybersecurity risks.