Google made Gemini 4 Carbon available on November 10, 2026 to developers, enterprises, scientists, and consumers, each with a different access route. One fact is worth checking before anything else, so we checked it first. On Google's own API pricing page, the Carbon row reads Not available on the Free tier. In the Gemini app, Carbon goes to Google AI Ultra subscribers, and eligible scientists can apply for six months of free access through Google's research program, subject to eligibility, capacity, and selection.
What Google released
On October 28, Google announced Gemini 4 Carbon and released it to cyber defenders through its Fairwind Program. On November 10, the company expanded access through paid developer, business, consumer, and scientific routes: the paid Gemini API tier in Google AI Studio and Android Studio, Gemini for business, and, for consumers, the Gemini app, AI Mode in Search, and Google Antigravity for Google AI Ultra subscribers.
Google's announcement calls Carbon the company's strongest model and positions it for complex, multi-step coding and debugging and open-ended scientific research. The company says Carbon uses more tokens than Argon to explore thoroughly, reason deeply, and run extended verification loops.
On benchmarks, Google reports Carbon ranks #1 on CursorBench 4.0, Perplexity's WANDR, and Zapier's AutomationBench, at more than 2x lower cost per task than the next best external model. On the Multimodal Knowledge Agent benchmark, it ranks #2, at more than 6x lower cost per task than the top model. These are benchmarks Google chose for its own announcement.
Why a beginner might care
The plan you're on decides whether you'll see Carbon at all. On Google's plans page, Google AI Plus and Google AI Pro include Gemini 4 Argon, while Google AI Ultra includes Carbon. In our November 10 check, a paid work account that isn't on Ultra did not show Carbon in the model picker, which is consistent with that plan distinction.
In the same announcement that opened Carbon up, Google says Argon remains the ideal choice for the majority of software engineering and knowledge work. Carbon is not presented as the default for every task. For many beginner tasks, the models you already have are enough.
The intro price runs through the end of the month. The API intro price is $2 per million input tokens and $10 per million output tokens through November 30, 2026. Starting December 1, both double to $4 and $20. Google's consumer plans page lists AI Ultra from $99.99 per month in the US on monthly billing.
What we checked on launch day
The model docs list it. Google's public API documentation shows gemini-4-carbon with a 1,048,576-token context window, up to 65,536 output tokens by default, and text, image, video, audio, and PDF as input types.
The pricing page is direct. Carbon reads Not available on the Free tier, with the paid tier at $2 and $10 per million tokens through November 30 and $4 and $20 starting December 1. The same note says every token the model processes is billed, including thinking tokens.
The app picker on a non-Ultra account shows no Carbon. We opened the Gemini app on November 10 with one paid Google work account that isn't on AI Ultra. The model picker listed Gemini 3.6 Flash, 3.8 Thinking, and 3.1 Pro, and no Carbon. That matches the plan map above.
We did not run the model itself. Our test setup holds no paid API tier and no AI Ultra plan, and the free tier excludes Carbon, so there was no path we could test end to end. The guide labels which steps come from our checks and which come from Google's documentation.
The honest catch
Carbon is built to use more tokens than Argon, and on the API, every processed token is billed, including thinking tokens. Google says that's how the model explores thoroughly, reasons deeply, and runs extended verification loops, so heavier token use can raise the cost of an API request. Google's consumer plans work on usage limits rather than per-token billing; heavier token use can still reach those limits sooner.
The intro price is temporary. From December 1, 2026, the API price doubles to $4 per million input tokens and $20 per million output tokens.
The benchmark rankings come from Google's announcement, and Google picked the benchmarks. On the Multimodal Knowledge Agent benchmark, Carbon ranks #2, not #1.
If your days are chat, search, and documents, Google's positioning keeps Argon as the model for most of that work. Carbon is positioned for complex, multi-step coding and debugging and open-ended scientific research.
The full walkthrough
Our study guide covers the plan requirements first, then a model-picker check you can run on your own account, the developer route into AI Studio, and a version stamp with everything we did and didn't test: Gemini 4 Carbon: who can use it now, and when to stick with Argon.
FAQ
Is Gemini 4 Carbon free?
No. The API pricing page reads Not available on the Free tier for Carbon. In the Gemini app, Carbon is rolling out to Google AI Ultra subscribers, and Google's plans page shows AI Plus and AI Pro include Argon instead. Eligible scientists can apply for six months of free access through Google's research program. Onboarding is rolling, and access is subject to eligibility, capacity, and selection.
What's the difference between Gemini 4 Carbon and Gemini 4 Argon?
Argon is Google's default recommendation for the majority of software engineering and knowledge work. Carbon is positioned above it for complex, multi-step coding and debugging and open-ended scientific research, and it uses more tokens than Argon to explore, reason, and verify. On Google's plans page, Plus and Pro include Argon while Ultra includes Carbon.
How much does Gemini 4 Carbon cost?
On the API, $2 per million input tokens and $10 per million output tokens through November 30, 2026, then $4 and $20 starting December 1. Cached input tokens are $0.10 per million through November 30, then $0.20. For consumers, Carbon comes with Google AI Ultra, listed from $99.99 per month in the US on monthly billing.
Can I use Gemini 4 Carbon in the free Gemini app?
Carbon is not included with the free API tier, and Google's announcement names Google AI Ultra subscribers as the consumer audience for Carbon in the Gemini app and AI Mode in Search. Google AI Plus and AI Pro list Argon instead. Our November 10 check of one paid, non-Ultra work account showed no Carbon in the model picker, and availability can roll out over time.
Which Gemini model should I use day to day?
Google says Argon is the ideal choice for the majority of software engineering and knowledge work, and free models are where many beginners start. Carbon is positioned for complex, multi-step coding, debugging, and open-ended scientific research, with longer token use billed on the API.
Want the plan-by-plan version? The guide includes the plan requirements, a model-picker check you can run on your own account, the developer route into AI Studio, and a record of what we did and did not test: learn.greatwithtech.com/guides/gemini-4-carbon.
More on this: What Is a Token in AI? The Unit Behind Every Price and Limit and Why Does AI Usage Run Out So Fast? How Subscription Limits Work.