Gemini 4 Carbon is now available for developers.

Google has introduced Gemini 4 Carbon, positioning it as the company’s strongest model with frontier performance in cybersecurity and extended task execution. The model is designed to handle complex, multi-step workflows, particularly in software engineering, full-stack web development, and scientific research. Its architecture supports sustained agentic workflows, enabling reliable performance over extended periods. Developers are encouraged to leverage its alignment and follow-through to streamline secure product development.

Gemini 4 Carbon is now accessible through the paid tier of the API in Google AI Studio and Android Studio, expanding its availability to a broader developer audience. Google AI Ultra subscribers gain access to the model in AI Studio and Google Antigravity, Google’s agent-first development platform. This integration aims to provide developers with tools for building and deploying advanced AI-driven solutions. The model’s capabilities are tailored to address real-world challenges in software and research environments.

The announcement highlights Carbon’s strong cybersecurity capabilities, positioning it as a tool for developers prioritizing secure product development. Its performance in long-horizon tasks ensures it can manage intricate, multi-step processes without compromising reliability. The model’s alignment features are designed to reduce errors and improve consistency in outputs, aligning with industry standards for security and efficiency.

Developers can begin using Gemini 4 Carbon immediately, with access points available in Google AI Studio, Android Studio, and Google Antigravity. The model’s availability in these platforms underscores Google’s commitment to integrating advanced AI tools into existing development workflows. Its focus on cybersecurity and complex problem-solving makes it a practical choice for teams aiming to enhance their development processes.