The latest in AI, every dayAI News
← AI News

October 30, 2026 · Khaleej Times

Google Launches Gemini 4 Carbon AI Model With Cybersecurity Focus

My take: Google unveiled Gemini 4 Carbon, a model focused on cybersecurity: it finds vulnerabilities, exploits them in controlled tests, and helps patch code, priced at $2 per million input tokens and $10 per million output tokens. For now it's only being tested by a small group of cybersecurity defenders inside Google's Fairwind program. It isn't available to the public.

Google says Carbon produces secure code with more than double the success rate of its previous model, Gemini 4 Argon, according to its own internal tests. It's worth being clear here: since the same company that builds the model ran that comparison, we will never know for certain whether the test is free of bias. Those numbers are worth reading with your own judgment, and it would be wise to wait for independent testing before treating them as definitive.

What is clear is where the industry is heading: increasingly specialized models for high-stakes tasks like cybersecurity, instead of one generic model for everything. If your business handles sensitive data or critical systems, are you thinking about how ready your infrastructure is for when these tools reach your industry?

Read at the source: Khaleej Times ↗

Want to use these tools? See the unbiased reviews or back to the news.