Google’s Gemini 4 Carbon is impressive—but not generally available yet

Google announced its new cyber-defense-focused frontier model October 28. The action today is to evaluate its published capabilities and access policy, not assume it can be added to an API stack.

Google introduced Gemini 4 Carbon on October 28 with strong claimed results on secure coding and multi-step agentic work. Google cites 88.2% on SEC-bench Pro, 54.8% on ExploitGym and 71% on CWE-bench. These are the company’s reported evaluation figures and should be compared with independent testing when access broadens.

Crucially, Carbon is being offered first to selected cyber defenders and trusted partners through the Fairwind Program. Google says broader developer, enterprise and consumer access will follow, but a general API launch is not confirmed today. Its announced introductory API pricing is $2 per million input tokens and $10 per million output tokens when access opens.

For a software consulting business, the relevant capability is finding, reproducing and patching subtle defects across large codebases. Save the official technical announcement and add Carbon to a future coding-model benchmark rather than budgeting it as an available production dependency.

Go to Source — [blog.google](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/) https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/