Back

Google's Gemini 4 Carbon Takes the Top Spot on the Artificial Analysis Intelligence Index

Google’s Gemini 4 Carbon has taken first place on the Artificial Analysis Intelligence Index v5, scoring 59 and showing strong reasoning. The model holds the top score in four of the index’s nine evaluations: HLE-Diamond, Terminal-Bench-Science, AA-Omniscience and AutomationBench-AA.

Strong in Coding, but Not Leading Everywhere

On the Artificial Analysis Coding Agent Index, Gemini 4 Carbon scores 59 when used with Antigravity CLI, three points behind Claude Code with Claude Opus 5.5 (62). It ranks second on both Terminal-Bench 4.0 (64%) and FrontierCode v1.1 (51%), behind Claude Opus 5.5.

In agentic knowledge work, it reaches 1609 Elo on AA-Briefcase, up 126 points from Gemini 4 Argon. That still trails Grok 4.7, Muse Wonder and Kimi K3.1 Pro.

Cost and Accuracy

At standard pricing, Gemini 4 Carbon costs $7.76 per Intelligence Index task, about 23% less than Claude Opus 5.5 ($10.04). It uses 123k output tokens per task, 29% fewer than Opus 5.5.

On AA-Omniscience v1.1, it scores 54, ahead of Claude Opus 5.5 (48). Accuracy rises to 58%, while the hallucination rate climbs from Argon’s 4% to 11%, still the second lowest of any model tested.

Availability and Pricing

Gemini 4 Carbon offers a 1M-token context window and accepts text, image, video and speech input, with text output. Standard pricing is $4 per 1M input tokens and $20 per 1M output tokens, and Google is offering a 50% launch discount with no confirmed end date.

The model is being rolled out to selected users and is not yet publicly available.

Artificial Analysis’s page for Gemini 4 Carbon (High), showing its intelligence, performance, cost and verbosity results


Source: Artificial Analysis
Originally Published: 28 October 2026

Contact usWhatsApp
EN