Google DeepMind Unveils Gemini 4 Carbon Frontier Model
Google DeepMind introduces Gemini 4 Carbon, a flagship model for cyber defense and coding, scoring 59 on the Artificial Analysis Intelligence Index and 88.2% on SEC-bench Pro.
Stock photo for illustration only, not from the actual event
- Google DeepMind introduces Gemini 4 Carbon, a new frontier AI model
- Engineered specifically for cyber defense and software coding tasks
- Achieves a score of 59 on the Artificial Analysis Intelligence Index
- Scores an impressive 88.2% on the SEC-bench Pro benchmark
Google DeepMind has officially announced the launch of Gemini 4 Carbon, a cutting-edge frontier model engineered to excel in cyber security and software development. This release marks a significant milestone in specialized artificial intelligence, tailored to tackle increasingly sophisticated digital threats and complex coding environments.
In benchmark evaluations, Gemini 4 Carbon scored 59 on the Artificial Analysis Intelligence Index. Furthermore, the model demonstrated exceptional proficiency on the SEC-bench Pro benchmark, securing an impressive score of 88.2% which underscores its advanced analytical capabilities.
Regarding initial availability, Google DeepMind has implemented strict access controls. Access is initially restricted to pre-screened and vetted cyber defenders, ensuring that this powerful technology is deployed responsibly and kept out of malicious hands.
Restricting initial access of Gemini 4 Carbon strictly to vetted security professionals highlights Google's commitment to responsible AI development. Frontier models with advanced security and coding capabilities carry dual-use risks, meaning they could potentially be exploited for cyber attacks if widely available without oversight. Controlled rollout strategies are becoming an industry standard for high-capability models.
Stock photo for illustration only, not from the actual event
As for the pricing structure, Google DeepMind has set the initial API access rates at $2 per million tokens for input and $10 per million tokens for output, reflecting the immense computational resources required to operate such advanced frontier intelligence models.
Source: MarkTechPost
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment