Google unveils Gemini 4 Carbon, its most powerful model yet
What's the story
Google has unveiled Gemini 4 Carbon, the latest addition to the Gemini series of artificial intelligence (AI) models. The new model is being touted as the strongest one yet, boasting cutting-edge cybersecurity and long-horizon capabilities. It builds on the success of its predecessor, Gemini 4 Argon, which was released just two weeks ago.
Enhanced features
Performance on external benchmarks
Gemini 4 Carbon ranks first on the latest version (v5.0) of the Artificial Analysis Intelligence Index, achieving a score of 59.
This shows major improvements in coding and agentic tasks, as well as key capabilities critical for cybersecurity defense.
The model has also performed exceptionally well on SEC-bench Pro (88.2%) for finding vulnerabilities, ExploitGym (54.8%) for exploitation, and CWE-bench (71.0%) for patching, according to external benchmarks.
Market introduction
Rollout to begin with select group of cyber defenders
The rollout of Gemini 4 Carbon will be gradual, starting with a select group of cyber defenders.
The process is part of the US government's voluntary pre-release model testing.
The model will be made available to other trusted partners through the Fairwind Program.
Once all feedback has been gathered and guardrails iterated, it will be made accessible to developers, businesses, and consumers via product surfaces as soon as possible.
Cost details
Pricing for Gemini 4 Carbon
Gemini 4 Carbon will be priced at the same introductory rates as Argon: $2 per million input tokens and $10 per million output tokens.
Cached input tokens will be offered at a 95% discount on the input token price.
The new model is designed to tackle complex, long-horizon engineering problems and has shown impressive performance in internal tests over multi-day runs.
Security features
Major improvements in generating secure code
Gemini 4 Carbon has made huge strides in cybersecurity.
It has achieved state-of-the-art performance across the entire cybersecurity defense lifecycle, from vulnerability detection to validation and remediation, acording to the official blog post.
The model has more than doubled the success rate for generating secure code over Argon in internal benchmarks.
It also performed over 10 times better than another lab's large frontier model when generating high-quality patches, according to the Google Chrome team.
Twitter Post
Gemini 4 Carbon is going to roll out gradually
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model.
— Sundar Pichai (@sundarpichai) October 28, 2026
Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys… pic.twitter.com/GN9LqZ7Zcf