Skip to content
Trending
BreakingAI1 min read

Gemini 4 Carbon access: live for AI Ultra, $2 per million input tokens until Nov 30

Gemini 4 Carbon is live for AI Ultra subscribers — $2/1M input tokens until Nov 30, then $4 — with Argon still available, 3.1 Pro gone, and a new benchmark showing Carbon leads in multimodal agentic reasoning.

By Model Card

The AI Desk · (3 hours ago)

Reported fromoutlet, credited1
Photo via 9to5Google

Gemini 4 Carbon is live in the Gemini app for AI Ultra subscribers — and it’s priced, positioned, and benchmarked as Google’s new top-tier reasoning engine. Not a replacement for Argon. Not a general-purpose upgrade. A specialist tool: $2 per million input tokens and $10 per million output tokens until November 30, then $4/$20 thereafter.

Available now in the Gemini app’s AI Mode and Antigravity interface. No rollout delay. No beta tag. Just live access, if you’re on AI Ultra.

Argon remains fully available for AI Pro and above, explicitly retained for “code & analysis” workloads. Carbon is slotted beside it under “Most complex work.” The model picker makes the split unambiguous. So does Google’s framing: Argon stays the default for software engineering and knowledge work. Carbon is for what comes after that, multi-step debugging across unfamiliar codebases, open-ended scientific hypothesis generation, architectural audits, legal document synthesis with strict schema compliance, or financial modeling where every cell in an output spreadsheet must be traceable and format-locked.

It’s not about speed. It’s about depth, fidelity, and chain-of-thought resilience across dense multimodal inputs: blueprints, annotated schematics, tabular data with embedded charts, PDFs with layered vector graphics. That’s why Google partnered with Artificial Analysis to launch the Multimodal Knowledge Agents (MKA) benchmark, 200 agentic tasks built around real-world professional workflows.

Carbon ranked #1. And used three times fewer tokens than the next best model. Antigravity’s /swarm preview shows how it’s meant to be used: one prompt spawning a coordinated team of agents, Carbon handling high-stakes reasoning roles, lighter models managing parallel verification, isolation, and execution. The Gemini app has dropped 3.1 Pro entirely.

This isn’t an expansion. It’s a reordering. Carbon sits at the apex, narrow, expensive, precise. Argon holds the broad middle.

And the pricing window ends November 30. After that, the cost doubles. AI Ultra subscribers get first access. Everyone else waits.

Reported from

How this story was made

Written by the Hitechreports desk from the reporting credited above, with facts attributed to their original publishers. We do not test devices ourselves; anything about performance, battery life or cameras comes from the outlets that did. Prices are as reported at the time of writing. Editorial policy · Report an error

What did you make of it?

Filed by The AI Desk

Models, assistants and the companies and chips behind them, reported from what was released and what was claimed, with the difference kept clear.

More from this desk →

Be the first to comment

Join the argument. No password, just your email or a passkey.

    Read next

    More AI →