Google introduces Gemini 4 Carbon model with cybersecurity focus
The new model is Google's latest frontier AI system, first offered to selected cyber defenders in a phased rollout.
By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .
This story is no longer being updated. It reflects reporting up to .
What happened
Google introduced Gemini 4 Carbon, a new Google DeepMind model that it describes as its strongest to date. The model brings improvements in cybersecurity, coding, reasoning, long-horizon agentic workflows and safety, and it is initially available to selected defenders through Google's Fairwind Program. A later updated Carbon checkpoint was spotted on Arena, where testers said it was faster, better at 3D and thought for longer.
Why it matters
Google says the model is its strongest yet, and its early access is aimed at cybersecurity defenders. One report says it lags behind Claude on benchmarks but offers better value.
Timeline
- Oct 28: Google announced Gemini 4 Carbon with cybersecurity benchmark scores and a phased rollout.
- Nov 1, 3:43 AM ET: An updated Carbon checkpoint surfaced on Arena, and testers said it was faster and better at 3D.
Coverage
More from Technology & AI
- China puts humanoid robots to work serving food and sorting parcels
- Newsom proposes sweeping AI safety rules for Washington
- Google releases Gemini 4 Neon AI model with million-token context
- Microsoft and Intel widen their cloud chip partnership
- Google DeepMind rolls out Genie 4 to select subscribers
- AI helped AMD shorten chip design work, OpenAI says