Google introduces Gemini 4 Argon for defensive cybersecurity

Alphabet, the parent company of Google, has introduced Gemini 4 Argon, a new artificial intelligence model built to handle tasks such as coding, research, and writing. Google highlights cybersecurity as a primary area of focus for the new model.

Controlled rollout through the Fairwind Program

Argon is not available to the general public. Instead, it is being rolled out to a select group of cyber partners through Google's Fairwind Program, which is the company's security initiative.

The model was trained specifically for defensive cyber work. According to Google, the system can autonomously find, validate, and patch critical software vulnerabilities.

Capabilities and internal use

Beyond cybersecurity, Argon is designed for coding and engineering tasks. Google states that its own staff have already used the model for daily work, including debugging and codebase migrations.

The model is also equipped to parse visuals, allowing it to analyze the contents of long videos and charts. In a blog post published on Wednesday, the company noted that Argon was built to sustain deep reasoning across complex, long-horizon workflows.

Benchmarks and competition

Google claims that Argon scored higher than competing models, including OpenAI's GPT-6 Astra and Anthropic's Fable and Opus models, across various artificial intelligence benchmarks. The company cited Vals, an artificial intelligence benchmarking startup, to support its claim that Argon leads the company's model index.