Google Introduces Gemini 4 Argon AI Model

Google has introduced Gemini 4 Argon, a new frontier artificial intelligence model designed to handle complex, long-duration tasks that require advanced reasoning. According to the company, the model is built for coding, multimodal understanding, and enterprise applications across sectors such as finance and law. It can also identify, verify, and fix software vulnerabilities, potentially helping cybersecurity professionals detect threats more effectively. Google said Argon helped uncover a security flaw in healthcare software used by hospitals worldwide, which could have exposed sensitive personal information. Given its powerful cybersecurity capabilities, the company is initially restricting access to trusted cyber defenders participating in its Fairwind Programme.

Gemini 4 Argon features an expanded output limit of one million tokens, compared with the previous limit of 64,000, allowing it to work through lengthy tasks in greater depth. Google says its engineers are using the model for debugging, large-scale code migrations, and algorithm design, while researchers are exploring applications in quantum computing and data-centre optimisation. The company also reported strong benchmark results in software engineering, financial research, legal work, business automation, and long-video understanding. Built-in safeguards aim to prevent misuse, prompt injection attacks, and attempts to escape controlled environments. Google plans to expand access gradually, beginning with trusted cybersecurity partners before making the model available to eligible paid API customers and Google AI Ultra subscribers.