Google releases Gemini 4 Carbon model for developers and enterprises
Google has launched Gemini 4 Carbon, describing it as its most capable model yet for cybersecurity tasks and long-running agentic work spanning coding, knowledge tasks, and scientific reasoning. The model is available via API in Google AI Studio and Gemini for business at $2 per million input tokens and $10 per million output tokens, and is rolling out to Google AI Ultra subscribers in Antigravity, the Gemini app, and AI Mode in Search. Google also previewed a new /swarm feature in Antigravity that coordinates multiple AI agents from a single prompt.
GoKawiil's interpretation of the reporting above, not reported fact.
Google claims Carbon ranks first on several independent benchmarks while costing less than half as much per task as rival models, which could pressure competitors on both performance and price. The distinction between Argon for everyday coding work and Carbon for the hardest reasoning tasks suggests Google is building a tiered model lineup aimed at different cost-performance needs rather than a single flagship product. The /swarm feature hints at a broader push toward multi-agent systems as the next stage of AI-assisted software development.
- Gemini 4 Carbon is now generally available via API and to Google AI Ultra subscribers
- Google says it leads on CursorBench 4.0, WANDR, and AutomationBench at over 2x lower cost than rivals
- A new /swarm feature in Antigravity lets Carbon coordinate teams of AI agents on complex projects
Source: blog.google — Koray Kavukcuoglu, 2026-11-10
Published there as: “Gemini 4 Carbon is now available”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.