Part of our AI models explained: ChatGPT, Claude, Gemini and the rest guide
Google is making Gemini 4 Carbon available to developers, enterprises and consumers, after releasing it about two weeks ago to a set of cybersecurity defenders through its Fairwind Program. The model is now in the Gemini API, in Gemini for business, and, for Google AI Ultra subscribers, in the Gemini app, AI Mode in Search, and Google Antigravity, at an introductory price of $2 per million input tokens and $10 per million output tokens through November 30, 2026, then $4 and $20. Google is also previewing /swarm in Antigravity, which it says turns a single prompt into a coordinated team of agents.

Who can use it, and what it costs
In a November 10 post, Koray Kavukcuoglu, Google DeepMind's chief AI architect, called Carbon the company's "strongest model" for cybersecurity work and long-horizon agentic workflows. That follows our October report on the Fairwind-only launch, when developers had no public API date.
Google says Carbon is available via the API in Google AI Studio and Gemini for business for all paying developers and enterprises, "at the same introductory price as Gemini 4 Argon." AI Ultra subscribers can use it in Antigravity, the Gemini app, and AI Mode in Search. The post also points developers to Android Studio, and enterprises to the model dropdown, Antigravity, or CodeMender. Cybersecurity teams can still apply to Fairwind for Carbon for cyber. The November 30 price change "applies to the Carbon model intended for public access only."
| Period | Input (per million tokens) | Output (per million tokens) |
|---|---|---|
| Through November 30, 2026 | $2 | $10 |
| After November 30, 2026 | $4 | $20 |
Prices are from Google's November 10 post.
A day after Anthropic's flagship drop
Carbon's public launch lands a day after Anthropic released Claude Fable 5.2 and Mythos 5.2 on November 9. Two frontier-model makers shipping new flagship systems a day apart is the story, not a verdict on either one. Google still describes Argon as "the ideal choice for the majority of software engineering and knowledge work," and Carbon as the model for harder, longer jobs that "use more tokens than Argon to explore thoroughly."
The /swarm preview is part of that pitch: Google says developers can assign Carbon to roles that need the deepest reasoning, such as architectural design, coordination and auditing, and pair it with faster models in parallel workspaces. Codema's AI Models hub tracks flagship launches and list prices as they land, and our AI models explained guide puts Carbon next to ChatGPT and Claude.
Google also reports its own benchmark results, including a #1 rank on CursorBench 4.0, Perplexity's WANDR and Zapier's AutomationBench, at "more than 2X lower cost per task compared to the next best external model." Those figures are Google's. We are not treating them as independent scores.
What to watch
- Whether the introductory $2/$10 price holds through November 30, and how the $4/$20 list is applied after that.
- How
/swarmbehaves in real projects, beyond the demos in Google's post. - Who actually gets access in the Gemini app and AI Mode, which Google limits to AI Ultra subscribers in this rollout.
FAQ
Who can use Gemini 4 Carbon now, and what does it cost?
Google is making it available to paying developers and enterprises via the Gemini API in Google AI Studio and Gemini for business, and to Google AI Ultra subscribers in the Gemini app, AI Mode in Search, and Google Antigravity. The introductory price is $2 per million input tokens and $10 per million output tokens through November 30, 2026, then $4 and $20.
How is Carbon different from Gemini 4 Argon?
Google still describes Argon as "the ideal choice for the majority of software engineering and knowledge work," and Carbon as the model for harder, longer jobs that "use more tokens than Argon to explore thoroughly." Carbon had been limited to cybersecurity defenders through the Fairwind Program before this public opening.
What is `/swarm` in Google Antigravity?
A preview Google says turns a single prompt into a coordinated team of agents. Developers can assign Carbon to roles that need the deepest reasoning, such as architectural design, coordination and auditing, and pair it with faster models in parallel workspaces.
Sources
Topics
John F.AI desk
John F. is a Codema.ai house byline: a pen name used by our editorial desk, not one individual journalist. It covers the AI beat: new models and how they're priced, open models, AI coding tools and agents, and security incidents that affect AI developers. Pieces under this name are written with AI assistance and reviewed by a human editor before they're published.
All stories by John F.


