Models · Updated 28 Oct, 10:31 pm IST
Google DeepMind announces Gemini 4 Carbon frontier model focused on cybersecurity
Why it matters for readers: If you follow AI progress, this is a claim about stronger security and coding performance and a cautious phased rollout.
- Gemini 4 Carbon is presented as DeepMind’s strongest model with state-of-the-art cybersecurity and long-horizon capabilities.12
- DeepMind reports Carbon ranks #1 on the Artificial Analysis Intelligence Index v5.0 with a score of 59.12
- The company cites top scores on security benchmarks: SEC-bench Pro (88.2%), ExploitGym (54.8%), and CWE-bench (71.0%) for vulnerability finding, exploitation, and patching respectively.12
- DeepMind says Carbon builds on Gemini 4 Argon (launched two weeks earlier) and will be made available this week to a subset of cyber defenders.12
- DeepMind is participating in the U.S. government’s voluntary process for pre-release model testing as part of its deliberate rollout.12
Get a brief like this every morning
Uzha reads hundreds of sources and gives you the stories that matter for your work, with every source linked. Free.
Get startedHow this story developed
28 Oct, 10:30 pm · Google DeepMind
Gemini 4 Carbon: Our next frontier model
28 Oct, 10:31 pm · Hacker News (100+ points)
Gemini 4 Carbon