Veriwire
← Back to the wire

Gemini 4 Carbon is the new #1 on the Artificial Analysis Intelligence Index v5

AchievementBenchmarkOct 28, 2026

Google's Gemini 4 Carbon ranks #1 on the Artificial Analysis Intelligence Index v5 with a score of 59, second only to Claude Opus 5.5 on the Coding Agent Index. It achieves lower Cost per Task than Claude Opus 5.5 at comparable intelligence. Like Gemini 4 Argon, it is rolling out to selected users and is not yet publicly available, with a 50% launch discount.

Why it matters

Google's Gemini 4 Carbon has taken the top position on Artificial Analysis's Intelligence Index v5, overtaking rivals on the strength of its complex reasoning, knowledge, and scientific task performance. The model posts the leading result on several of the index's individual evaluations, though it does not lead in coding and knowledge work. On the Coding Agent Index, it ranks second, trailing only Claude Opus 5.5, a notable shift given Google's historically weaker showing in that area. Its agentic knowledge work improves over Gemini 4 Argon but still lags several competitors. A key part of the story is cost: Gemini 4 Carbon achieves comparable overall intelligence to Claude Opus 5.5 while costing less per task, aided by lower token usage per task. Like Gemini 4 Argon, it is rolling out only to selected users and is not yet publicly available, with Google applying a launch discount of unconfirmed duration. Artificial Analysis charts use standard pricing.

Receipt № 26411 source · awaiting confirmation ◐

Evidence

1source· awaiting independent confirmation

No score is assigned. Sources and their independence are shown in the citation chain below.

Citation chain · 1 source

01high
artificialanalysis.aiOfficialhttps://artificialanalysis.ai/articles/gemini-4-carbon-new-number-1This source is the index publisher itself, providing the benchmark results, cost comparisons, and rollout details for Gemini 4 Carbon directly from Artificial Analysis.
PRIMARY
GoogleCompanyArtificial AnalysisCompanyClaude Opus 5.5ModelGemini 4 ArgonModelGemini 4 CarbonModel

Related

Oct 28Our AI Agent Executed the Same Successful SQL Query 254 TimesOct 27Can we trust LLMs to judge whether a research idea is novel?Oct 25Anthropic co-founder Jack Clark moves his prediction for self-improving AI forward to 2027
Canonical: https://artificialanalysis.ai/articles/gemini-4-carbon-new-number-1