# Google Launches Gemini 4 Carbon Model for Cybersecurity Google debuts high-performance Gemini 4 Carbon with aggressive pricing and enhanced security capabilities. - **Source:** [www.theregister.com - Articles](https://brifing.si/source/2c34ea50-d341-4a10-925d-12c012a9ad10) - **Published:** 2026-10-29 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/56ff1da0-af3d-4596-b640-7df92db99364 ## Rapid Model Iteration Strategy Just two weeks after releasing Gemini 4 Argon, Google announced the Gemini 4 Carbon model. This new iteration offers superior performance while maintaining the same pricing structure temporarily. Google aims to solidify its position at the AI frontier by rapidly deploying upgraded versions to stay ahead of competitors like Anthropic and OpenAI. ## Strategic Pricing and Competition Gemini 4 Carbon is currently priced at $2 per million input tokens and $10 per million output tokens. This introductory rate matches the previous Argon model. Once the promotional period ends, prices will double to $4 and $20 respectively, aligning directly with the cost structure of Anthropic's Claude Opus 5.5 model. ## Cybersecurity Focus and Performance Koray Kavukcuoglu highlighted Carbon's specific strengths in cybersecurity and long-running tasks. Internal benchmarks indicate a 2x success rate increase in generating secure code compared to Argon. Furthermore, the Chrome team reported that Carbon performed 10x better than other frontier models when generating high-quality patches for various software development needs. ## Benchmark and Leaderboard Rankings Carbon leads in SEC-bench Pro, ExploitGym, and CWE-bench, though it holds fourth place on WebDev Arena. It currently ranks second behind Claude Opus 5.5 on the Arena.ai leaderboard for Best Overall Agent. Additionally, it tops the Artificial Analysis Intelligence Index v5 while securing second place on the Coding Agent Index. ## Deployment and Access Plans Google is releasing Gemini 4 Carbon this week to a select group of cybersecurity defenders. Broader availability is planned through the company's Fairwind access program as soon as possible. This phased rollout allows Google to refine the model's security applications before providing general access to the wider developer community and enterprise users. ## Scale of Model Usage Sundar Pichai confirmed that Google's first-party model APIs are now processing nearly 32 billion tokens per minute. This volume represents a fourfold increase compared to the previous year. Such massive scale demonstrates the infrastructure capacity Google leverages to support its growing ecosystem of AI-driven products and services for global users. ## Financial Performance and Revenue Google reported a half-trillion dollar run rate in its Q3 2026 earnings, significantly outpacing competitors. This revenue figure is approximately ten times higher than that of OpenAI or Anthropic. Despite the intense competition in model performance, Google maintains a distinct financial advantage that supports its ongoing, aggressive investment in AI development. ## Market Positioning and Outlook The launch of Carbon signals Google's intent to dominate the cybersecurity AI niche. By emphasizing secure code generation and high-quality patching, the company targets critical enterprise needs. Future success depends on maintaining this performance lead while balancing the eventual price increase against the established market presence of rival AI model providers. ## Sources covering this story - https://www.theregister.com/ai-and-ml/2026/10/29/googles-gemini-4-carbon-looks-like-a-diamond-in-the-rough/5306149 — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._