Google Cloud’s public SKUs catalog now lists Gemini 4 Carbon at $4.00 per million input tokens and $20.00 per million output tokens for online predictions on the global endpoint. That is twice the $2 and $10 introductory price Google gave when it announced Carbon on 28 October, and the same as Anthropic’s list price for Claude Opus 5.5, by our arithmetic. Google has not announced general availability, and Carbon is not on its Vertex AI pricing page.
TestingCatalog, an X account that tracks unreleased AI features, posted at 08:25 UTC on 10 November that Carbon “is now listed in the GCP SKUs catalog”. We read the catalog ourselves at about 09:35 UTC and counted 225 Carbon rows: 90 for Gemini 4 Carbon and 90 for a Gemini 4 Carbon Cyber Permissive version under Vertex AI, plus 45 under the Gemini API.
Business Pill · Usage-Based Pricing
An evergreen Business Engineer explainer on usage-based pricing: paying for work, not for seats. It explains the general idea only and does not discuss Gemini 4 Carbon.
The key insight: As we read it, the catalog puts a second number next to the one Google announced: Gemini 4 Carbon’s online rate is $4.00 and $20.00 per million tokens, twice the $2 and $10 introductory price of 28 October and level with Claude Opus 5.5’s list price, by our arithmetic.
The Vertex AI rows price Gemini 4 Carbon by input type (text, image, video and audio) and by serving mode. Online Predictions, the standard on-demand mode, are $4.00 per million input tokens and $20.00 per million text output tokens, with cached input at $0.20. Batch, Flex, Offpeak and Deferred Predictions are $2.00 and $10.00, half the online rate, with cached input at $0.10.
Priority Predictions are $7.20 and $36.00, 1.8 times the online rate, by our arithmetic. Rows labelled Regional are $4.40 and $22.00, 10% above the global rows. The Cyber Permissive rows carry the same prices as the matching standard rows. The 45 Gemini API rows follow the same pattern: $4.00 input and $20.00 output, $2.00 and $10.00 for batch and flex, $7.20 and $36.00 for priority.
The catalog shows descriptions, SKU IDs and prices only. It gives no launch date and does not say who can use the model.
Output price per million tokens, with input / output labels: Gemini 4 Carbon at Google’s 28 October introductory price ($2 / $10) and at the Google Cloud SKU catalog’s online rate for global text ($4 / $20), against Anthropic’s list prices for Claude Opus 5.5 ($4 / $20) and Claude Fable 5.2 ($8 / $40). Data: Google Cloud SKUs page and Anthropic pricing page, read 10 November 2026; Google launch post, 28 October 2026.
What Was Posted on X
GOOGLE ๐ฅ: Gemini 4 Carbon appears to be prepared for release on Google Cloud and is now listed in the GCP SKUs catalog.
The “gemini-4-carbon” model slug may already be responsive as well.
Carbon day ๐
h/t @AiBattle_
— ๐จ AI News | TestingCatalog (@testingcatalog) Nov 10 08:25
TestingCatalog posted at 08:25 UTC on 10 November that Gemini 4 Carbon โappears to be prepared for release on Google Cloud and is now listed in the GCP SKUs catalogโ, and that the model slug โmay already be responsive as wellโ. We read Googleโs SKUs page ourselves and give its prices above; we did not call the model.
How the Price Compares With 28 October
Google’s launch post of 28 October said “Gemini 4 Carbon will launch at the same introductory price” as Argon at “$2 per million input tokens and $10 per million output tokens”, “with cached input tokens priced at 95% off input token price”. The catalog’s $0.20 cached rate is 95% below its $4.00 input rate, by our arithmetic.
Argon shows how the two figures can sit side by side. The same catalog lists Gemini 4 Argon’s online text rows at $4.00 and $20.00. Google’s Vertex AI pricing page shows Argon at $2.00 and $10.00 “through November 30, 2026” and $4.00 and $20.00 “Starting December 1, 2026”, noting: “Promotional pricing provided through 50% credits back on net spend on select models within a given period”.
Our 28 October piece put Carbon at half of Opus 5.5 on the introductory price. At the catalog’s online rate, Carbon matches the $4 and $20 that Anthropic’s pricing page lists for Opus 5.5, and is half the $8 and $40 listed for Claude Fable 5.2, which Anthropic released on 9 November, by our arithmetic.
Where Carbon Stands on Google’s Own Pages
When we checked between 09:29 and 09:45 UTC on 10 November, Carbon did not appear on the Vertex AI pricing page, the Vertex AI models page, the Vertex AI release notes (latest entry 6 November), the Gemini API models page or the Gemini API changelog. A search of recent X posts from Google, Google DeepMind, Google Cloud, Logan Kilpatrick and other Google accounts found no Carbon announcement.
Google said on 28 October that it would make Carbon “available to a subset of cyber defenders” and “gradually phase access to additional trusted partners through our Fairwind Program”. TestingCatalog wrote that the gemini-4-carbon model slug “may already be responsive as well”; we did not call the model. The Gemini web app menu has also shown an unannounced next-generation agent for enterprise, as our 8 November piece reported.
The Structural Read
As we read it, Argon is the template: its catalog rows also say $4.00 and $20.00, while Google’s pricing page shows the $2.00 and $10.00 Argon rate as a promotion paid through 50% credits until 30 November.
As we read it, 225 rows across Vertex AI and the Gemini API, including a separate Cyber Permissive version, is a full price grid rather than a single placeholder entry.
Google, launch post, 28 October 2026
“Gemini 4 Carbon will launch at the same introductory price”
Three Implications
FOR BUYERS COMPARING FRONTIER MODELS At the catalog’s online rate, Carbon costs the same per token as Opus 5.5 and half as much as Fable 5.2, by our arithmetic; the $2 and $10 introductory price Google announced is half of Opus 5.5’s list price.
FOR BATCH WORKLOADS Batch, Flex, Offpeak and Deferred rows are $2.00 and $10.00 in the catalog, the same as the introductory online price Google announced.
FOR GOOGLE CLOUD The catalog lists Carbon under both Vertex AI and the Gemini API, with Priority rows at $7.20 and $36.00, 1.8 times the online rate, by our arithmetic.
The framework says of the token: “It is simultaneously the unit of cognition the model produces, the unit of compute the data center serves, the unit of price the lab charges, and the unit of value the enterprise extracts.”
As we read it, a SKU catalog is where the unit of price becomes concrete: the same Carbon token carries different prices by serving mode and region. The framework predates Google’s Carbon listing and does not discuss it.
What Is Not Established
Whether Google applies the $2 and $10 introductory price to Carbon through credits, as it does for Argon until 30 November, is not stated on any page we read; the pricing page has no Carbon rows. For Argon, the page says the promotion does not apply to the Cyber Permissive version, and it says nothing yet about Carbon’s.
A catalog listing is not an availability notice: Google’s launch post gave no date for wider access, and we did not test whether the model responds. Rates at Opus 5.5’s level matter for heavy users, as our piece on Claude Ultra’s limits shows, but the effective price paid depends on credits and modes the catalog does not show.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
Google Cloud’s SKU catalog lists Gemini 4 Carbon at $4.00 and $20.00 per million tokens online, twice the $2 and $10 introductory price Google announced on 28 October and level with Claude Opus 5.5, by our arithmetic. Google has not announced availability, and its pricing page does not list Carbon or say how the introductory price applies to it, as it does for Argon.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Prices and SKU names are from Google Cloud’s public SKUs page and Vertex AI pricing page, read on 10 November 2026; the introductory price is from Google’s 28 October launch post; Opus 5.5 and Fable 5.2 prices are from Anthropic’s pricing page. TestingCatalog’s post was the tip. We did not call the model. All ratios and comparisons are our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
