# Google's Gemini 4 Carbon costs more per task than Argon, rising in price from 30 November

> Google released Gemini 4 Carbon via API and to Ultra subscribers, with a price rise due 30 November.

*Google's strongest model undercuts rivals on task cost in its own benchmarks, but its API price doubles from 30 November.*

By Bez · InsidersFeed
Canonical: https://insidersfeed.com/news/google-s-gemini-4-carbon-costs-more-per-task-than-argon-rising-in-price-from-30

Google has released Gemini 4 Carbon, which it calls its strongest model, through the API and to Google AI Ultra subscribers. The introductory price is $2 per million input tokens and $10 per million output tokens, the same as Argon. From 30 November 2026 it rises to $4 input and $20 output per million tokens, doubling the price for developers who build on it.

Google's post, written by Koray Kavukcuoglu, SVP, Google DeepMind and Chief AI Architect, says Carbon is built for long-horizon agentic workflows, complex knowledge work, software engineering and scientific reasoning. The company recommends Argon for most software engineering and knowledge work, noting that Carbon uses more tokens to reason deeply and run extended verification loops. Kavukcuoglu said it is "our strongest model with frontier performance in cybersecurity capabilities."

Google [claims](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-carbon-now-available/) Carbon ranks first 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. It ranks second on the Multimodal Knowledge Agent benchmark at more than 6X lower average cost per task than the top model. The company also says Carbon blocked 99% of attack trajectories in simulated tests, and that its per-attempt breach rate fell to 0.8% from 3.1% on Argon. These figures are Google's own and have not been independently verified.

Carbon is available in Google AI Studio and Gemini for business for paying developers and enterprises. It is rolling out to Ultra subscribers in Google Antigravity, the Gemini app and AI Mode in Search. A new `/swarm` preview in Antigravity turns one prompt into a team of agents tracked in real time. Carbon handles high-reasoning roles such as architecture, coordination and auditing, while faster models run in parallel workspaces with verifiers checking every change. Scientists can apply for six months of free access to Argon and Carbon in Antigravity's highest usage tier, subject to eligibility and capacity.

## The Insider view

The price change matters more than the launch. A developer who builds on Carbon at $2 and $10 today will pay double from 30 November, so anyone pricing agent products around the introductory rate should model the higher figure now. Google is pitching cost per task rather than cost per token, which puts the claim on the workload rather than the rate card. Those comparisons come from Google's own benchmarks, so buyers should run their own tasks before switching.

## Key takeaways

- Introductory API price is $2 input and $10 output per million tokens, rising to $4 and $20 from 30 November.
- Google claims more than 2X lower cost per task than the next best external model on three benchmarks.
- Carbon leads on Google's cybersecurity benchmarks and is in preview as a multi-agent team tool.

## Sources

- [Gemini 4 Carbon now available](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-carbon-now-available/) — Google, 2026-11-10
