Google launches Gemini 4 Carbon with a ‘significant leap’ in cybersecurity
The Verge Emma Roth ● Covered by 6 sources
Google says Gemini 4 Carbon is a new AI model with big gains in coding and cybersecurity. Right now it's limited to a small set of defense firms, but Google is pushing it wider later.
Based on reporting by The Verge, Emma Roth — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Google has pushed out Gemini 4 Carbon, its newest frontier model, and it’s claiming a noticeable jump in coding, agentic work, and cybersecurity. The company’s own framing is pretty blunt: this one is supposed to be faster and sharper where the stakes are highest.
Carbon also lands at the top of the Artificial Analysis Intelligence Index v5.0 benchmark, with Claude Opus 5.5 and GPT-6 Astra behind it. Google says the model improves on Argon, which it released earlier this month. That’s a short gap, even by AI standards, and it suggests the company is moving quickly from one iteration to the next.
For now, though, access is tight. Carbon is only available to a select group of cybersecurity defense companies. Google says it will eventually open the model to other partners in its Fairwind Program.
The company is pitching Carbon as proof of how quickly its models are advancing, especially in security-focused work. And if the benchmark claim holds up outside the lab, that’s the part the industry will care about most: not the branding, but whether the model can actually help defend systems without becoming another flashy demo.
My take — AI-written commentary, not fact-checked reporting
Google is doing what big AI labs always do when they feel heat: put up a benchmark win and call it progress. The real tell is the limited rollout to cybersecurity firms, because that’s where models can be useful without the company having to answer every messy question at once. Pretty classic move: sell the future, distribute the risk.
Read more about this at: The Verge