← Back to GPN · All articles

Google's Gemini 4 Carbon Is Now Available, Built for Deep Reasoning and Cybersecurity

OFFICIAL Written by GPN · Nov 10, 2026, 3:00 PM ET · · AI

Google has made Gemini 4 Carbon available to developers, enterprises, scientists and consumers. Carbon is the company's strongest model yet, built for frontier performance in cybersecurity and long-horizon agentic work across knowledge work, software engineering and scientific reasoning.

Carbon follows the recent launch of Gemini 4 Argon. Google positions Argon as the right choice for most software engineering and knowledge work, while Carbon is aimed at tasks that test the limits of a model's intelligence — complex, multi-step coding and debugging, and open-ended scientific research — using more tokens to explore thoroughly, reason deeply and run extended verification loops.

On independent benchmarks, Carbon ranks first on CursorBench 4.0, Perplexity's WANDR and Zapier's AutomationBench, at more than twice lower cost per task than the next best external model. Google is also previewing /swarm in Google Antigravity, which turns a single prompt into a coordinated team of agents, with Carbon assigned to the roles that demand the deepest reasoning, such as architectural design, coordination and auditing.

Carbon is available through the API in Google AI Studio and Gemini for business for all paying developers and enterprises, at the same introductory price as Argon — $2 per million input tokens and $10 per million output tokens. It is also rolling out to Google AI Ultra subscribers in Google Antigravity, the Gemini app and AI Mode in Search.

Cybersecurity is Carbon's standout area. Through the Fairwind Program, it leads on benchmarks for finding, exploiting and patching vulnerabilities. In Google's autonomous Project Fortify system it hit 100% recall on scanning and 95% precision on triage, and Wiz's Scan for Good program used Carbon to uncover zero-day vulnerabilities in under two hours — work Google says compresses weeks of research into a single afternoon.

Carbon is also showing strong results in science. Researchers used it alongside other models in AlphaProof Code to formally verify the code behind Signal's cryptographic protocol, a task that historically takes expert teams years. Google is expanding its program giving 10,000 scientists access to its frontier models in Antigravity free for six months, with Carbon now included.

On safety, Google says Carbon blocked 99% of attack trajectories in simulated cyberattack testing, achieved a 0% prompt injection attack success rate on the Gray Swan benchmark, and passed 92% of tasks on its 375-task agentic honesty benchmark, up from 70% for Argon.

Facts gathered from

• Google

More on this

Share