News in short
- Gemini 4 Carbon is now available to paying developers, enterprise customers and Google AI Ultra subscribers.
- Google positions Gemini 4 Carbon as its most capable model for cybersecurity and long-running agentic workflows.
- The model costs $2 per million input tokens and $10 per million output tokens through the paid Gemini API tier.
- Google Antigravity is previewing a /swarm feature that coordinates multiple AI agents for complex development tasks.
Google has expanded access to Gemini 4 Carbon, its advanced AI model built for complex reasoning and multi-step work. The model is now available to paying developers and enterprise users. Google AI Ultra subscribers can also use it across select products, including the Gemini app and AI Mode in Search.
Gemini 4 Carbon follows the launch of Gemini 4 Argon last month. Google says Carbon can spend more time reasoning, checking its work and handling demanding tasks. The company is targeting software engineering, scientific reasoning, cybersecurity and long-running agentic workflows.
What is Gemini 4 Carbon and what can it do?
Gemini 4 Carbon is Google’s latest high-capability AI model for tasks that need deeper analysis and repeated verification. According to Google, it is designed for complex knowledge work, software engineering and scientific reasoning. It can also support agentic workflows that involve several connected steps rather than a single prompt.
Google says Gemini 4 Carbon ranks first on three independent benchmarks: CursorBench 4.0, Perplexity’s WANDR and Zapier’s AutomationBench. The company also claims that Carbon delivers more than twice the cost efficiency per task compared with the next-best external model across these benchmarks. These are Google’s reported results and should be read in that context.
Google Antigravity gets a new team-agent feature
Google is also previewing a new /swarm feature in Google Antigravity. It can turn one prompt into a coordinated team of AI agents. Developers can monitor these agents in real time as they work across separate workspaces.
Developers can assign Gemini 4 Carbon to tasks that need deeper reasoning. These may include software architecture, coordination and code audits. Faster models can handle parallel tasks, helping teams divide complex projects into smaller workloads.
Gemini 4 Carbon focuses on cybersecurity
Cybersecurity is another major focus. Google is offering Gemini 4 Carbon for cyber through its Fairwind Program. The model can inspect software, identify vulnerabilities and trace how separate flaws could combine into larger attack paths. Google also says it can generate verified patches to address security issues.
Gemini 4 Carbon availability and API pricing
Developers can access Gemini 4 Carbon through the Gemini API in Google AI Studio and Android Studio. It is also available to enterprise customers through Gemini for Business, Google Antigravity and CodeMender. Google AI Ultra subscribers can access it through Google Antigravity, the Gemini app and AI Mode in Search.
Google has kept the API pricing in line with Gemini 4 Argon. The paid Gemini API tier costs $2 (around ₹187) per million input tokens and $10 (around ₹936) per million output tokens. Actual billing will depend on token usage and the applicable plan.
The expansion gives developers and businesses wider access to Google’s model for advanced coding, research and security tasks. For context, read our earlier report on Gemini 4 Argon and its features and pricing. You can also check Google’s official Gemini 4 Carbon announcement for further details.
