Gemini 4 Argon launches for developers and paid users
Google has expanded access to Gemini 4 Argon, its new flagship AI model built for software development, scientific research, cybersecurity, and complex knowledge work. The model is now available to developers, business customers, and subscribers to Google’s paid AI plans.
This marks a broader commercial launch rather than another product announcement. When Google introduced Gemini 4 Argon on September 30, 2026, access was initially limited to selected cybersecurity defenders. Google subsequently released the model to developers and businesses worldwide on October 13, according to Reuters.
Who can access Gemini 4 Argon
Developers and organizations can use Argon through Google’s AI application development tools. Google AI Pro and AI Ultra subscribers can also access the model through the Gemini app and AI Mode in Google Search.
Availability may still vary by region, account type, and product. Access through the API does not necessarily mean that every Argon-powered feature will appear in the Gemini app or Search at the same time.
Some capabilities are restricted to higher-priced plans. Argon integration in Google’s Antigravity coding environment, for example, is initially limited to AI Ultra subscribers.
What Gemini 4 Argon is designed to do
Google describes Argon as a model for long-running, multi-step workflows. Instead of generating a single response, it is designed to plan a task, use external tools, inspect its own output, and continue working toward a result.
Its intended applications include:
writing, testing, and modernizing software;
reviewing legal and financial material;
conducting scientific research;
detecting and fixing software vulnerabilities;
working with large collections of company knowledge;
powering autonomous AI agents.
Google says the model supports an output limit of up to one million tokens. In practical terms, that could allow it to produce or process substantially larger bodies of work without dividing a task into numerous separate sessions.
Google has also published benchmark and internal productivity results for Argon. These figures should be treated as company claims: real-world performance will depend on the tools, prompts, workloads, and evaluation methods used.
Gemini 4 Argon API pricing
The introductory API rate is $2 per million input tokens and $10 per million output tokens. Google says cached input tokens receive a 95% discount.
The launch pricing is temporary. Reuters reports that Google plans to double both rates after November 30. The final cost for developers will depend on prompt size, generated output, cached context, and the number of steps performed by an AI agent.
A major focus on cybersecurity
Cybersecurity is one of the most prominent use cases for Argon. Google says the model can find, validate, and help patch critical software vulnerabilities.
The company plans to expand its Fairwind Program to several hundred additional partners. Vetted cybersecurity defenders can receive a specialized version of Argon without some of the cyber-related restrictions applied to the standard model. This is controlled access and is not intended for general subscribers.
Google is also offering 10,000 scientists six months of free access to the highest Antigravity tier, giving selected researchers an opportunity to use Argon for complex scientific workflows.
What the launch means for users
For Google AI Pro and Ultra subscribers, Argon introduces a more capable model inside products they may already use. Developers gain another API option for agentic systems, while businesses can apply it to software, documents, research, and internal data.
Gemini 4 Argon is not being released as a universally available free model. Access is tied to paid subscriptions, business products, or API billing, while some advanced features remain exclusive to AI Ultra subscribers and vetted Google partners.