Google has made Gemini 4 Argon, its newest frontier artificial intelligence model, globally available to developers and enterprises through the Gemini API, Google AI Studio and the Gemini platform. Koray Kavukcuoglu, senior vice president at Google DeepMind and Google's chief AI architect, announced the wider release on October 13, two weeks after the model was first unveiled on September 30 with access limited to approved cybersecurity defenders.
Introductory pricing is $2 per million input tokens and $10 per million output tokens, with thinking tokens counted as output. Google's pricing page shows these rates apply through November 30, 2026, and double from December 1 to $4 and $20. Cached input costs $0.10 per million tokens until then, and the batch tier is priced at half the standard rate. Argon is not offered on the free tier of the API.

Beyond the API, Google AI Ultra subscribers can use the model in Google Antigravity, the company's agent-first development platform, while Ultra and Pro subscribers can use it in the Gemini app and in AI Mode in Search. Developers can also build with it in Android Studio. Google says the model supports a unified input and output limit of 1 million tokens through the Gemini API.
Google positions Argon for software engineering, enterprise knowledge work and cybersecurity defense. On the independent Vals Index, the model ranked first with 68.90% as of October 12, ahead of Anthropic's Claude Sonnet 5.5 at 67.04% and Claude Opus 5.5 at 66.97%, VentureBeat reported. In Google's own tests, Argon scored 87% on the Embodied Reasoning benchmark and held prompt injection attacks to a 0.7% success rate across 15 adaptive attempts.
Inside Google, more than 100,000 employees use Argon each month through an internal version of Antigravity. The company says teams using it increased agentic code submissions by more than 35% and cut code rollbacks by nearly a third. Google is also expanding its Fairwind Program, which gives cybersecurity defenders a version of the model without cyber guardrails, to hundreds of additional partners.
