Google Launches Gemini 4 Carbon, Outperforming Claude Opus 5.5 in AI Benchmarks
November 4, 2026 - Google has introduced Gemini 4 Carbon, a frontier AI model that has taken the top position on Artificial Analysis’ Intelligence Index V5, narrowly outperforming Anthropic’s Claude Opus 5.5.
The model also recorded the highest score on Artificial Analysis’ AA-Omniscience benchmark, which evaluates factual knowledge and hallucination tendencies. Google is initially restricting access to selected cybersecurity organizations through its Fairwing Programme.
Gemini 4 Carbon Performance and Benchmarks
Gemini 4 Carbon scored 59 on the Intelligence Index V5, one point ahead of Claude Opus 5.5 at 58 and two points ahead of OpenAI’s GPT-6 Astra at 57.
On the AA-Omniscience V1.1 evaluation, Carbon scored 54, compared with 58 for Claude Opus 5.5. Its reported hallucination rate was 11%, although Gemini 4 Argon performed better on this specific measure at 4%.
The model also supports complex reasoning, coding, and agentic workflows. However, Claude Opus 5.5 retains an advantage on certain programming evaluations, showing that Carbon does not lead across every category.
Pricing and Availability
Google is initially providing Gemini 4 Carbon to selected cybersecurity organizations through its Fairwind programme. Broader access is expected to expand through trusted partners, although Google has not announced a general release date.
The introductory API pricing is $2 per million input tokens and $10 per million output tokens, matching the introductory rates announced for Gemini 4 Argon. Standard pricing is listed at $4 per million input tokens and $20 per million output tokens.
Sources:
https://mezha.net/eng/news/mezha-b6066273f3735f1e9d13f168_google-s_gemini_4/
https://artificialanalysis.ai/models/comparisons/gemini-4-carbon-vs-claude-opus-5-5
