Google has officially announced the release of Gemini 4 Carbon, its latest frontier artificial intelligence model. This new system is specifically tuned to improve performance in coding, scientific tasks, and full-stack web development. The company states that the model excels in multimodal understanding and agentic workflows.
According to Google, Gemini 4 Carbon has secured the top position on version 5.0 of the Artificial Analysis Intelligence Index. The model outperforms most other available systems in general benchmarks. However, the primary focus of this release is on cybersecurity capabilities.
Security and Code Generation Benchmarks
Google claims that Gemini 4 Carbon significantly outperforms its predecessor, Gemini 4 Argon, in generating secure code. Internal benchmarks indicate that the success rate for creating secure code has more than doubled. The company also reported that the model performed more than 10 times better than another large frontier model when generating high-quality security patches.
The model was tested against several external benchmarks to verify its security posture. These tests included SEC-bench Pro for vulnerability detection, ExploitGym for exploitation scenarios, and CWE-bench for patching effectiveness. Google notes that the model builds upon the same safety and security stack as Argon but includes further improvements in alignment.
Release Strategy and Pricing
Google has chosen to follow the US government’s voluntary review process for this release. As a result, the model will not be available to the general public immediately. Instead, it will be rolled out this week to a specific subset of cyber defenders. This controlled release allows for further monitoring of the model’s behavior in real-world security scenarios.
When the model becomes available to the broader market, it will carry a specific pricing structure. The cost is set at $2 per million input tokens and $10 per million output tokens. Google also offers a discount for cached input tokens, which will be priced at 95% off the standard input token price.
The announcement follows the recent rollout of Gemini 4 Argon, which became available to everyday users two weeks ago. The rapid succession of these models highlights the fast pace of development in the AI sector. Google continues to refine its frontier models to address specific industry needs, with this latest iteration targeting the critical field of digital security.
Source: Android Authority
