# Google DeepMind unveils Gemini 4 Carbon, its strongest frontier model

published: 2026-10-28T18:05:41.917Z
topic: infrastructure
breaking: true
source: https://deepmind.google/blog/gemini-4-carbon-our-next-frontier-model/
url: https://aiagentnews.io/stories/google-deepmind-unveils-gemini-4-carbon-its-strongest-frontier-model

Google DeepMind has announced Gemini 4 Carbon, a frontier model it says is its strongest yet, with a focus on cybersecurity and long-horizon agentic work. Developer and consumer access is not yet open, with early access limited to cyber defenders and selected trusted partners.

Google DeepMind has announced Gemini 4 Carbon, which it describes as its strongest model, with state-of-the-art cybersecurity and long-horizon capabilities. Carbon ranks first on the Artificial Analysis Intelligence Index v5.0 with a score of 59. Google reports 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching.

Access is staged. From this week, Carbon is available to a subset of cyber defenders, and more trusted partners will be added gradually through the Fairwind Program. Google says it is still gathering feedback and refining guardrails before a wider release to developers, businesses and consumers, and that Carbon is part of the U.S. government's voluntary pre-release testing process.

The model launches at an introductory price of $2 per million input tokens and $10 per million output tokens, the same as Gemini 4 Argon. After the introductory period, pricing rises to $4 and $20 per million tokens, and cached input tokens cost 95% less than standard input.

## Why it matters if you build agents

Teams building agents for long multi-step coding or security work can plan to evaluate Carbon once it opens to them, since Google says it is designed for tasks that run for days. The reported gains are mostly vendor benchmarks and internal tests, so results should be checked on your own workloads. Budget for the higher post-introductory token prices when estimating costs.