Google Launches Gemini 4 Carbon, Its Second Major AI Model in a Month

Google unveiled Gemini 4 Carbon, which it said is significantly more capable than the model it released less than a month ago, as the company steps up efforts to contend with the leading frontier labs.

The back-to-back launches come after nearly a year where Google didn’t release a new flagship frontier model, ceding ground to rival labs OpenAI and Anthropic.

The new model is priced the same as Gemini 4 Argon, announced on Sept. 30 and released widely earlier this month, with an introductory price of $2 per million input tokens and $10 per million output tokens, the company said on Wednesday. That will double after the teaser rate ends.

Google says Carbon is a “significant leap in cybersecurity capabilities,” and is nearly two times better at generating secure code than Argon. Carbon’s release comes on the heels of a cybersecurity crisis as rogue AI agents have exploited both corporate and government vulnerabilities.

As with Argon, Google said it will release Carbon to a subset of cybersecurity professionals and then to partners in its Fairwind program. It also said it was actively engaged in the U.S. government’s program for pre-release model testing.