Model launches

AI-generated text

Google unveils Gemini 4 Carbon model focused on cybersecurity and software tasks

Google announced Gemini 4 Carbon, a top-tier AI model optimized for software development and cybersecurity workloads.

Google unveils Gemini 4 Carbon model focused on cybersecurity and software tasks

Google, a subsidiary of Alphabet, has unveiled its latest high-end AI model, Gemini 4 Carbon. According to the company, the new model is specifically aimed at strengthening performance in software development and cybersecurity use cases, positioning Google against competitors such as OpenAI and Anthropic.

Timing and product positioning

The Gemini 4 Carbon announcement follows the unveiling of Gemini 4 Argon by two weeks. With these rapid successive releases, Google is expanding the range of products it offers to developers and enterprise customers.

Designed use cases

Google says Gemini 4 Carbon delivers strong performance on a variety of cybersecurity tasks, including discovery and exploitation of vulnerabilities as well as remediation of security flaws. The model was designed to handle complex, long-running engineering and programming workloads.

According to Google's internal measurements, Carbon writes secure code more than twice as efficiently as the recently announced Gemini 4 Argon.

Access and safety testing

Initially, the Gemini 4 Carbon will be made available to a limited group of cybersecurity professionals through Google's Fairwind program. Google will continue testing the model's built-in safety constraints during this rollout before broader availability.

Pricing

Google stated that Gemini 4 Carbon carries the same introductory pricing as Argon: $2 per one million input tokens and $10 per one million output tokens.

Why this matters

The release signals Google’s intent to compete in the enterprise and security-focused model market by tailoring offerings to developer and corporate needs. Emphasizing cybersecurity capabilities reflects market demand for models that can assist with secure code generation and vulnerability management.

Additional notes

This article was prepared based on reporting that includes Reuters. The cover image is illustrative and credited to Getty Images. An AI assistant contributed to preparing this piece; the final content was edited and verified by a journalist.

Related event

The Portfolio AI & Digital Transformation conference will take place on November 26; topics at the event overlap with the AI applications and digital transformation issues discussed here.