Google has unveiled Gemini 4 Argon, a new flagship artificial intelligence model built to handle complex and lengthy tasks in areas such as coding, research, enterprise work and cybersecurity.
Google DeepMind announced the model on September 30, calling Argon its latest frontier AI system for software engineering, enterprise knowledge work and cyber defence. According to the company, thousands of Google employees are already using the model for specialised coding, research and writing tasks.
One of the major improvements in Gemini 4 Argon is its ability to produce up to 1 million tokens of output. This is a significant increase from the previous 64,000-token limit. Google says the larger capacity allows Argon to tackle complicated multi-step problems and lengthy tasks within a single workflow.
Cybersecurity is also a key area of focus for the new model. Google says Argon can independently identify, verify and patch critical software vulnerabilities. The model is initially being offered to selected cybersecurity organisations through the company’s Fairwind Program.
Read Also: Karachi launches Rs14bn dumlottee water supply project
Google is taking a phased approach to the launch due to Argon’s advanced cybersecurity capabilities. The company said it is working with early testers as well as the US government’s voluntary pre-release process to improve safety measures before making the model more widely available.
Argon is also being used at Google for major software engineering projects, including codebase migrations and optimisation. The company says its researchers have applied the model to quantum computing projects, while engineers are using it for debugging and large-scale code migrations.
Google has announced an introductory API price of $2 per million input tokens and $10 per million output tokens. Cached input tokens will be available at a significantly lower price. Wider access for developers, businesses and consumers is expected following additional safety testing, although Google has not provided a specific date for the public release.
The launch is part of Google’s continued push to strengthen its position in the rapidly growing AI market. Gemini is increasingly being used across both consumer and professional applications, while independent reports have positioned Argon as a competitor to the latest frontier models from OpenAI and Anthropic.