Home / AI News / Gemini 4 Carbon

Gemini 4 Carbon: Google's Strongest Model for Agentic Work, Coding and Research

Google released Gemini 4 Carbon, its strongest model for long, multi-step agentic work, with first-place results on coding, search and business automation benchmarks and a new /swarm multi-agent preview.

Spark5X · · Google

Google made Gemini 4 Carbon available on November 10, 2026, calling it its strongest model yet for long, multi-step agentic work across knowledge work, software engineering and scientific reasoning. Carbon is available to paying developers and enterprises through Google AI Studio and Gemini for business, and is rolling out to Google AI Ultra subscribers in Google Antigravity, the Gemini app and AI Mode in Search.

Key takeaways
  • Gemini 4 Carbon is Google's new frontier model, built for tasks that push a model's limits, while Gemini 4 Argon remains Google's recommended choice for most everyday coding and knowledge work.
  • Google says Carbon ranks first on CursorBench 4.0, Perplexity's WANDR and Zapier's AutomationBench, at less than half the cost per task of the next best external model.
  • Google Antigravity gains a /swarm preview that turns one prompt into a coordinated team of agents you can follow in real time.
  • For businesses, Carbon produces spreadsheets, presentations and reports that follow strict formats, and builds and checks full web applications.

1. What is Gemini 4 Carbon?

Gemini 4 Carbon is Google's most capable Gemini 4 model, designed for complex, long-running work such as multi-step coding and debugging, dense document analysis and open-ended research. Google positions it above Gemini 4 Argon, which it released shortly before: Argon stays the right choice for the majority of software engineering and knowledge work, while Carbon uses more reasoning to explore thoroughly, think deeply and run extended verification loops on the hardest tasks.

Before the general release, Google gave security teams early access to Carbon through its Fairwind Program, where it helped find and patch software vulnerabilities, including in widely used open-source code.

WhereWho gets Gemini 4 Carbon
Google AI Studio (API)All paying developers
Gemini for businessEnterprises on paid plans
Google AntigravityRolling out to Google AI Ultra subscribers, with the /swarm preview
Gemini app and AI Mode in SearchRolling out to Google AI Ultra subscribers

2. How does Gemini 4 Carbon perform on real work?

Google reports that Gemini 4 Carbon ranks first on three benchmarks built around real workflows — CursorBench 4.0 for coding, Perplexity's WANDR for agentic search, and Zapier's AutomationBench for business automation — at less than half the cost per task of the next best external model. Google also partnered with Artificial Analysis on a new Multimodal Knowledge Agent (MKA) benchmark, which tests how well agents reason over dense files containing charts, tables and diagrams. Carbon ranks second there, at less than a sixth of the top model's average cost per task.

In pre-release testing on finance work, Google says Carbon matched the best model tested on Hebbia's end-to-end financial agent evaluation and ranked first of 23 frontier model configurations on Balyasny Asset Management's General Financial Intelligence benchmark.

3. What is /swarm in Google Antigravity?

/swarm is a preview feature in Google Antigravity that turns a single prompt into a coordinated team of AI agents you can track in real time. Developers can give Carbon the roles that need the deepest reasoning — architecture, coordination and auditing — and pair it with faster, lighter models working in parallel, isolated workspaces, with separate verifier agents testing every change.

Google's demonstrations include a team of Carbon agents building an interactive 3D reconstruction of an office from a set of public engineering drawings, and a browser game built end to end, from 3D assets to music, from a single brief.

4. What can Gemini 4 Carbon do for knowledge work and web development?

For knowledge work, Gemini 4 Carbon reads across large sets of dense documents and data, combines what it finds into grounded insights, and produces spreadsheets, presentations and reports that follow strict schemas and formatting. Google highlights legal, finance and enterprise operations as early fits.

For web development, Carbon runs build-and-verify loops: it writes the code, recovers from errors, and checks both the code and the rendered interface to deliver complete, working applications. Google also reports stronger visual design. In one demonstration, Carbon reimagined the Google Merchandise Store in two prompts, proposing five art directions and generating new product imagery.

5. What does Gemini 4 Carbon mean for science and research?

Google says Gemini 4 Carbon shows strongly improved reasoning in physics, mathematics and life sciences, and is already accelerating research inside and outside Google. Drug toxicity research company Axiom Bio now rates Carbon its leading model for translational science, and Google researchers used it to help mathematically prove security properties of real cryptographic code in weeks rather than years. Google has also extended its program giving 10,000 scientists free access to its latest models so that participants can now use Carbon.

6. What does Gemini 4 Carbon mean for businesses?

For businesses, Gemini 4 Carbon makes it more practical to hand AI longer, multi-step jobs: research and reporting across many documents, automation that spans several tools, and complete web builds that check their own work. Its first-place AutomationBench result is directly relevant to teams automating sales and operations workflows, and its arrival in AI Mode in Search is another reason to make sure your site is easy for AI answers to understand and cite.

  • Automation: multi-agent setups like /swarm point to automations that plan, act and verify. Our AI automation agency team builds these workflows with a person approving what reaches customers.
  • AI search: as more answers come from AI Mode, structure your pages for citation; see our AI SEO agency services.
  • Planning: not sure where models like Carbon fit your business? Our AI consulting services rank where AI pays off first, and the AI News hub tracks each new release.
Sources. This article is based on the company's own announcement:

Frequently asked questions

What is Gemini 4 Carbon?

Gemini 4 Carbon is Google's strongest Gemini 4 model, released on November 10, 2026 for long, multi-step agentic work across knowledge work, software engineering and scientific reasoning. Google recommends Gemini 4 Argon for most everyday tasks and Carbon for the hardest ones.

Where can I use Gemini 4 Carbon?

Gemini 4 Carbon is available through the API in Google AI Studio and Gemini for business for paying developers and enterprises, and is rolling out to Google AI Ultra subscribers in Google Antigravity, the Gemini app and AI Mode in Search.

What is /swarm in Google Antigravity?

/swarm is a preview feature in Google Antigravity that turns one prompt into a coordinated team of AI agents you can track in real time, with Carbon taking the roles that need the deepest reasoning and lighter models working in parallel.

How does Gemini 4 Carbon compare with other models?

Google reports that Carbon ranks first on CursorBench 4.0, Perplexity's WANDR and Zapier's AutomationBench, at less than half the cost per task of the next best external model.

More AI news

All AI news →

Get started
Put this to work
Talk to us about using new AI tools in your marketing and growth systems.
    0/1000
    Prefer to pick a time? Book a call →
    🔒 We'll only use your details to reply. See our Privacy Policy.
    Keep reading

    Related AI news

    Request a quoteBook a call