Breaking

Google DeepMind Launches Gemini 4 Carbon, A Frontier Model Built for Cybersecurity and Long-Horizon AI Workflows

October 28, 2026 at 13:06 EDT

Foundation Models
AI Agents
Security

Google DeepMind introduced Gemini 4 Carbon on October 28, 2026, positioning it as the company's strongest frontier model yet, engineered specifically for sustained agentic workflows and advanced cybersecurity defense. The model represents a significant evolution in Google's AI capabilities, with particular strengths in vulnerability detection, code generation security, and multi-step engineering tasks that unfold across extended operational horizons.

October 2026 · Google DeepMind

Gemini 4 Carbon: Frontier Model for Security & Long-Horizon AI

Google launches its strongest model yet, engineered for autonomous AI agents, multi-step cybersecurity tasks, and sustained agentic workflows with state-of-the-art security performance.

88.2%

Vulnerability Detection (SEC-bench Pro)

Secure Code Generation Improvement

10x

Better Patch Generation (vs Competitors)

$2/1M

Input Token Pricing (Introductory)

Cybersecurity Benchmark Leadership

88.2%

SEC-bench Pro
(Vulnerabilities)

54.8%

ExploitGym
(Validation)

71.0%

CWE-bench
(Patching)

✓ Strengths

Multi-step agentic workflows
Enhanced security reasoning
10x better patch generation
Strong 3D scene generation

⚠ Trade-offs

2–4x slower inference
Limited public availability (early access only)
Weaker UI/frontend generation
No confirmed public release date

Rollout Plan

Week of Oct 28: Limited access to cyber defenders via Fairwind Program

Gradual expansion: Additional trusted partners onboarded

Public launch: Developer, business, and consumer access "as soon as possible"

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.