Google Launches Gemini 4 Argon Model for Developers and Enterprises
Google has released its new Gemini 4 Argon model globally, offering advanced capabilities for complex coding, legal, and cybersecurity tasks at a set introductory price.
Key points
- Gemini 4 Argon is now available via API and specific Google AI subscription tiers.
- The model ranks first on the Vals Index for performance in economic sectors.
- Google reports using the model to refactor Chrome code and optimize open-source projects.
Google has made Gemini 4 Argon, its latest frontier artificial intelligence model, globally available today for developers, enterprises, scientists, and consumers. The release provides access through the API in AI Studio and the Gemini platform, as well as to Google AI Ultra subscribers in Google Antigravity and Ultra or Pro subscribers in the Gemini app and AI Mode in Search.
The model is priced at an introductory rate of $2 per million input tokens and $10 per million output tokens, with cached input tokens receiving a 95 percent discount. According to Google, Argon excels at deep work in complex, real-world tasks across coding, finance, legal, cybersecurity, and science. It ranks number one on the Vals Index, which measures model performance across critical economic sectors including finance, coding, legal, and tax work.
Engineering and Research Applications
Google states that teams have used Argon to solve major engineering problems, such as large codebase refactoring, and to find and patch critical cybersecurity vulnerabilities. The company also cites breakthroughs in quantum computing, nuclear fusion, and life sciences. At Google, software engineers have reportedly shipped high-quality code faster, with Argon in Antigravity helping teams increase agentic code submissions by more than 35 percent while reducing code rollbacks by nearly a third.
Specific examples include the Chrome team using Argon to separate core webpage handling from a legacy folder of browser features within Chromium’s 40 million-line codebase. This architectural refactor, completed in weeks rather than months, aims to allow live webpages to move between tabs and panels without reloading, which Google says will eliminate a category of browser crashes.
Open-Source and Performance Optimizations
Google researchers also used Argon to optimize the open-source xTB quantum chemistry package, making it up to 2.3 times faster at predicting molecular structures. Additionally, the team behind Google’s open-source Agent Development Kit (ADK) used Argon to cut in-memory session prep time by 40 percent and speed up cold imports by 21 percent.
The model also helped achieve a 2.85 times end-to-end inference speedup on Google’s open Gemma models via vLLM on an A100 GPU. This optimization accelerated throughput from approximately 37,900 to over 108,000 tokens per second on Gemma 4 E2B, with Google noting 100 percent token parity and zero quality drop in a process that typically takes months.
What should you do?
Security and technology professionals should monitor the availability of Gemini 4 Argon through their preferred access channels, such as the API or Google AI subscriptions. Developers may want to evaluate the model’s performance on complex multi-step tasks, particularly in coding and cybersecurity, using the provided benchmarks and case studies. Organizations should assess the cost implications of the introductory pricing structure, noting the significant discount for cached input tokens.
What to do and how to stay safe: Google
- Review the pricing structure for input and output tokens to estimate costs for enterprise usage.
- Evaluate the model’s performance on specific benchmarks like the Vals Index for relevant domains.
- Monitor official channels for updates on API access and subscription tier requirements.
- Assess internal workflows for potential integration with agentic coding tools like Antigravity.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
How much does Gemini 4 Argon cost?
The introductory price is $2 per million input tokens and $10 per million output tokens, with cached input tokens at 95 percent off.
Who can access Gemini 4 Argon?
It is available via API for developers and enterprises, and to Google AI Ultra and Pro subscribers in the Gemini app and Search.
What benchmarks does Gemini 4 Argon lead?
It ranks number one on the Vals Index, which measures performance in finance, coding, legal, and tax work.
