Google launches Gemini 4 Argon for coding and cybersecurity
Alphabet has launched Gemini 4 Argon, a new Google AI model designed for tasks including coding, research and writing. Google is initially making it available to a limited group of cyber partners through Fairwind, its security initiative.
The model was trained for defensive cybersecurity work. Google says Argon can autonomously identify, validate and patch critical software vulnerabilities. The company also describes it as useful for engineering tasks, saying its employees have used it for debugging and codebase migrations. Argon can additionally process visual information, including material from lengthy videos and charts.
In a blog post, Google said the model is designed for complex, extended workflows. The company also claimed that Argon outperformed OpenAI’s GPT-6 Astra and Anthropic’s Fable and Opus on several AI benchmarks. Google cited Vals, an AI benchmarking startup, which placed Argon at the top of its model index. The launch follows recent releases from OpenAI and Anthropic as major AI companies compete to promote increasingly capable systems. Google said in August that its Gemini app had more than a billion monthly users, matching a figure later reported by OpenAI for ChatGPT.
Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.