Gemini 4 Carbon: our next frontier model
Google presents Gemini 4 Carbon as its strongest model so far, with notable gains in cybersecurity and long multi-step tasks. It is first limited to a small group of cyber defenders and trusted partners, with wider access planned later.
What the article says
Google announced Gemini 4 Carbon on October 28, 2026, in a post by Koray Kavukcuoglu, calling it its strongest model, two weeks after Gemini 4 Argon. The model is aimed at cybersecurity and long-horizon, agentic work. Rollout is staged: starting the week of the announcement, a subset of cyber defenders has access, with more trusted partners added through the Fairwind Program. Google is taking part in the U.S. government's voluntary pre-release testing and says broad availability to developers, businesses and consumers is planned soon.
Google cites a score of 59 on the Artificial Analysis Intelligence Index v5.0 (ranked first), 88.2 percent on SEC-bench Pro, 54.8 percent on ExploitGym and 71.0 percent on CWE-bench, plus an 89-point Elo gain on WebDev Arena (1768, fourth place). The introductory price is $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 afterwards. These results come from Google's own testing, and the page does not describe independent verification. For now, most users cannot access the model.
Read the official article AnnuIA profile
Summary written by AnnuIA in its own words from the official page of Google DeepMind. The content and rights belong to the publisher: please read the original article.