Gemini 4 Carbon: Google Conquers the Intelligence Frontier
Spearheaded by Koray Kavukcuoglu, Gemini 4 Carbon tops the Intelligence Index v5, signaling Google’s decisive comeback against rivals Anthropic and OpenAI
TL;DR
- Gemini 4 Carbon ranks number one on the Artificial Analysis Intelligence Index v5 with a score of 59.
- The model excels in complex reasoning, knowledge, and scientific tasks, and shows gains in coding and knowledge work.
- Gemini 4 Carbon is designed for complex, long-horizon engineering problems and sustained agentic workflows.
- It ranks second on coding agents, behind Anthropic's Claude Opus 5.5, and leads AA-Omniscience v1.1 with a low hallucination rate.
- The model supports text, image, speech, and video input, with a 1 million token context window.
- Gemini 4 Carbon is currently being rolled out to selected users and is not yet publicly available.
- Google's previous model, Gemini 4 Argon, scored 53 on the Intelligence Index, tying with OpenAI's GPT-6 Astra.