Google Opens Gemini 4 Carbon to Paid API Users and Ultra Subscribers
Gemini 4 Carbon costs $2 per million input tokens and $10 per million output tokens on the paid Gemini API until 30 November 2026. From 1 December 2026 the price rises to $4 for input and $20 for output. Google made the model available to paying developers and enterprises on 10 November 2026. It is not on the API free tier. Consumers reach it through Google AI Ultra, not Google AI Pro, which still gets the earlier Gemini 4 Argon model. For a business that automates work through the API, the per-token rate is the first number to check, but the final bill depends on how many tokens each task uses.
In short
- Gemini 4 Carbon is on the paid Gemini API from 10 November 2026, at $2 input and $10 output per million tokens.
- The API rate doubles to $4 and $20 on 1 December 2026, so tests and budgets should use the higher figure.
- Carbon is not on the free API tier, and Google AI Pro subscribers get Gemini 4 Argon instead.
- Per-token price is only part of the cost; the number of tokens a task uses sets the final bill.
- Google still recommends Argon for most routine work, so compare both models on real tasks before 1 December.
What happened
On 10 November 2026, Google opened Gemini 4 Carbon to all paying developers and enterprises through the Gemini API in Google AI Studio and Gemini for business, according to Google's own announcement. The same day, Google began rolling the model out to Google AI Ultra subscribers in the Gemini app, AI Mode in Search and Antigravity, its AI coding platform. Several outlets report that Google first announced the model in late October 2026, when only a group of cybersecurity defenders could use it.
Google set the introductory API price at the same level as its earlier Gemini 4 Argon model. Android Authority and other outlets confirm that both rates double on 1 December 2026.
Who can use it, and at what price
Access depends on how a business or person buys in. The table below sets out the routes reported in Google's announcement and in coverage from Android Authority and Search Engine Land.
| Route | Who it is for | Carbon access and price |
|---|---|---|
| Gemini API (paid tier) | Developers, via Google AI Studio and Android Studio | $2 input and $10 output per million tokens until 30 November 2026; $4 and $20 from 1 December |
| Gemini for business | Enterprises | Same API rates; the model can be chosen from the model menu |
| Google AI Ultra | Individual subscribers in supported countries | Included in the plan, with daily usage limits in AI Mode; Ultra starts at about $100 a month in the US |
| Google AI Pro | Individual subscribers | Not included; Gemini 4 Argon is the model on offer, and Google has announced no date for Carbon |
| Gemini API free tier | Anyone without a paid account | Not available |
The limits are worth stating once. Carbon is not on the free tier, so a business cannot trial it on the free quota. In AI Mode, Carbon is limited to Ultra subscribers in certain countries, and Google has not listed them in its launch post, according to Martincid's report.
Why it matters for businesses that automate work
Many automated workflows, such as reading invoices, matching payments or drafting reports, call a model through the API. The price per million tokens sets the running cost of each call. Google describes Carbon for demanding work across legal, finance and enterprise operations, and says it spends more tokens than Argon while it reasons through a problem. A higher token count means the per-token rate is only part of the cost of a task.
Google also continues to recommend Argon for most routine work, according to Android Authority. That recommendation matters for a business that runs simple, high-volume tasks. Paying for the stronger model on every step may not be the right choice.
What changes in practice
For a business of any size, the change is a decision about which model each automated task should use before 1 December. A workable approach looks like this:
- List every automated task that calls a Gemini model through the API, and note which model each one uses.
- Run a sample of real invoices, reports or messages through Carbon and through Argon, and compare the accuracy and the number of tokens used.
- Calculate the monthly cost at $4 and $20 per million tokens, not only at the introductory rate.
- Keep the cheaper model for routine steps, and reserve Carbon for the few steps where the extra reasoning changes the result.
Staff who use the Gemini app will see different options depending on their plan. Google AI Pro subscribers can use Argon, while Carbon needs Ultra. A business that buys seats for staff should check which plan each person holds before assuming they have the same model.
What to watch next
The first date to watch is 1 December 2026, when the API rate doubles. Google's announcement gives no timeline for Carbon on lower-priced plans, so Pro access is unconfirmed. Watch for an update to the list of countries where Carbon is available in AI Mode, and for independent tests of Carbon against Argon on routine business tasks. Google's own benchmark claims, which include results on a workflow automation test, are not independent checks.
Sources
How AiStaffo would automate this
AiStaffo would start by listing each automated task that calls a Gemini model, such as invoice reading, payment matching or weekly reporting, and recording which model and how many tokens each step uses. The automated workflow would run the same sample of real documents through Carbon and Argon and log the results side by side, so the cost at $4 and $20 per million tokens is visible before 1 December. The routine steps would stay on the cheaper model, with Carbon reserved for the few steps that need it. The owner still checks the accuracy of the outputs, approves the model choice and decides the monthly budget. To measure what this means for your own figures, book a free automation audit.
Questions people ask
How much does Gemini 4 Carbon cost on the API?
Can I use Gemini 4 Carbon for free?
Which Google plan gives access to Gemini 4 Carbon in the app?
Is Gemini 4 Carbon the right model for routine business tasks?
Why might the cost of a task be higher than the token price suggests?
Book a free automation audit
Thirty minutes. We look at one process you run every week and tell you exactly what an AI worker would take off your desk, and what it would not.
