Google unveils Gemini 4 Carbon for cybersecurity — Channel NewsAsia

Google has unveiled Gemini 4 Carbon, a new artificial intelligence model for cybersecurity, programming and engineering development tasks. As Channel NewsAsia reports, the model was launched amid competition between Google, OpenAI and Anthropic in developing AI systems for code writing and cybersecurity.

Model tasks

Google said that Gemini 4 Carbon delivers the highest results in cybersecurity tasks, including vulnerability detection, exploitation and software patching. The model is also designed to carry out complex and long-running engineering and programming tasks.

According to Google's internal tests, Carbon generates secure code with a success rate more than twice as high as Gemini 4 Argon. The company introduced Argon two weeks before Carbon's launch as part of expanding the Gemini lineup for developers and corporate clients.

More current news is available on the UA.News Telegram channel Telegram.

Limited access and price

Initially, Gemini 4 Carbon will be available to some cybersecurity specialists through the Google Fairwind Program. The company continues to test the model's safeguards.

Gemini 4 Carbon will cost $2 per million input tokens and $10 per million output tokens. This matches the initial price of Gemini 4 Argon.