Google has made Gemini 4 Carbon broadly available, according to a November 10 post on Google's blog by Koray Kavukcuoglu of Google DeepMind. Two weeks earlier, the model had been released only to a set of cybersecurity defenders through Google's Fairwind Program.
Developers can use Carbon through the paid tier of the Gemini API in Google AI Studio and Android Studio. Enterprises get it through Gemini for business, Google Antigravity and CodeMender. On the consumer side, it is rolling out to Google AI Ultra subscribers in the Gemini app, AI Mode in Search and Antigravity. Scientists can apply for six months of free access through Antigravity's highest usage tier, and defenders can still get "Carbon for cyber" through Fairwind.
The introductory price is $2 per million input tokens and $10 per million output tokens, the same as Gemini 4 Argon. When the introductory period ends on November 30, 2026, the price rises to $4 per million input tokens and $20 per million output tokens.
Google says Carbon ranks first on CursorBench 4.0, Perplexity's WANDR and Zapier's AutomationBench, at less than half the cost per task of the next-best external model. It also says Carbon leads on the SEC-bench Pro, ExploitGym and CWE-bench v1 security benchmarks. On safety, Google reports that Carbon blocked 99% of attack trajectories from an adaptive red-teaming agent, and that the automated cyberattack query success rate in Gray Swan testing fell to 2.17%, compared with 9.86% for Argon.
Google positions Carbon above Argon for the hardest work, such as complex multi-step coding, debugging and open-ended research, and still recommends Argon for most software engineering and knowledge work. For teams building multi-agent systems, Google highlights a new /swarm feature in Antigravity, now in preview, which starts a coordinated team of agents from a single prompt. Teams trying Carbon at the introductory rate should plan for the price to double at the end of November.