Google has launched Gemini 4 Argon, a powerful AI model for complex tasks in various fields, with special features for safety and accessibility.
Google has officially launched its latest AI model, Gemini 4 Argon, designed to tackle complex engineering, scientific, and analytical challenges.
Gemini 4️⃣ Argon ✨ is now rolling out globally. Our frontier model is built for complex, real-world tasks across coding, finance, legal, cybersecurity, and science. Check out how you can use Argon in your favorite Google product 🧵
— Google AI (@GoogleAI) October 13, 2026
This model features a context window capable of handling 1 million tokens for both input and output data. According to developers, Argon achieved top ranking in the Vals Index benchmark, which assesses the performance of neural networks in programming, finance, taxation, and law.
Google has already tested the model on internal projects, including refactoring the architecture of the Chrome browser by separating web page functions from an outdated codebase of 6 million lines. In robotics, the model generated code for the humanoid robot Apollo, enabling it to accurately replicate throwing motions.
Developers have implemented measures to protect against misuse, claiming to have reduced the likelihood of successful indirect prompt injection attacks to 0.7%. A real-time monitoring system oversees the model's reasoning chain, blocking attempts to bypass rules or gain unauthorized access to data.
Gemini 4 Argon is accessible via API in Google AI Studio and to subscribers of Google AI Ultra and Pro plans through the web version and Google's proprietary search. The starting cost for using the API is $2 per million input tokens and $10 per million output tokens, with rates set to increase on December 1 to $4 and $20, respectively. Additionally, Google will offer free access to the model for six months to 10,000 scientists and researchers.
In October, Google also introduced the image generation model Nano Banana 2.1 and launched an AI platform for creating games from text descriptions.
